This is a digital copy of a book that was preserved for generations on library shelves before it was carefully scanned by Google as part of a project
to make the world's books discoverable online.
It has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject
to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books
are our gateways to the past, representing a wealth of history, culture and knowledge that's often difficult to discover.
Marks, notations and other marginalia present in the original volume will appear in this file - a reminder of this book's long journey from the
publisher to a library and finally to you.
Usage guidelines
Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the
public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing this resource, we have taken steps to
prevent abuse by commercial parties, including placing technical restrictions on automated querying.
We also ask that you:
+ Make non-commercial use of the files We designed Google Book Search for use by individuals, and we request that you use these files for
personal, non-commercial purposes.
+ Refrain from automated querying Do not send automated queries of any sort to Google's system: If you are conducting research on machine
translation, optical character recognition or other areas where access to a large amount of text is helpful, please contact us. We encourage the
use of public domain materials for these purposes and may be able to help.
+ Maintain attribution The Google "watermark" you see on each file is essential for informing people about this project and helping them find
additional materials through Google Book Search. Please do not remove it.
+ Keep it legal Whatever your use, remember that you are responsible for ensuring that what you are doing is legal. Do not assume that just
because we believe a book is in the public domain for users in the United States, that the work is also in the public domain for users in other
countries. Whether a book is still in copyright varies from country to country, and we can't offer guidance on whether any specific use of
any specific book is allowed. Please do not assume that a book's appearance in Google Book Search means it can be used in any manner
anywhere in the world. Copyright infringement liability can be quite severe.
About Google Book Search
Google's mission is to organize the world's information and to make it universally accessible and useful. Google Book Search helps readers
discover the world's books while helping authors and publishers reach new audiences. You can search through the full text of this book on the web
at|http : //books . google . com/
of tbe
mnivereiti? of Misconef n
Google
11
II
f!
n
r
ii
! 'i
■
Digitized by VjOOQ IC
Digitized by VjOOQIC
ELECTRIC SMELTING AND REFINING.
Digitized by VjOOQ IC
GRIFFIN'S SCIENTIFIC TEXT-BOOKS.
Sbcond Edition, Thoroughly Revised and Enlarged. With Nnmeroiu niiutrations
and Tliree Folding Plates. Price 218.
ELECTRIC SMELTING AND REFINING. A Practical Manual of the
Extraction and Treatment of Metali by Electrical Methods. Being the "Elektro-
Metallurgie" of Dr. W. Borchers. Translated from the Third German Edition,
with Additions, by Walter G. McMillan, F.I.C, F.C.S.
"Comjprehensive and authoritative . . . Full of valuable information."— 7Ae
Eledrietan.
Second Edition, Revised, Enlarged, and In part Re-written. With Additional
Sections on Modern Theorie.s of Electrolysis, Costs, &c. Price lOs. 6d.
A TREATISE ON ELECTRO-METALLURGY. By Walter G.
McMillan, F.I.C., F.C.S., Secretary to the Institution of Electrical Engineers ; late
Lecturer in Metallurgy at Mason College, Birmingham. With Numerous Illustra-
tions. Large Crown 8vo. Cloth.
"This excellent treatise, ... one of the best and most complete manuals
hitherto published on Electro-Metallurgy."— j^^trieoj Review.
Large 8vo. Handsome Cloth. With Illustrations. Price 128. 6d. net.
METALLURGICAL ANALYSIS AND ASSAYING. A Three Years^
Course for StudenU of Schools of Mines. By W. A. Macleod, B.A., B.Sc., A.O.S.M.
(N.Z.). Formerly Assist. -Director, Thames School of Mines (N.Z.), and Lecturer in
Chemistry, University of Tasmania ; Director of Queensland Government Sch«x)l of
Mines, Charters Towers: and Chas. Walker, F.C.S. , Formerly Assist -Demon-
strator in Chemistry, Sydney University ; Lecturer in Chemistry and Metallurgy,
Charters Towers School of Mines.
'* The publication of this volume tends to prove that the teaching of metallurgical
analysis and assaying in Australia rests in competent hands."— Aa(ur«.
With 15 Folding Plates and 160 Illustrations in the Text. 2l8. net.
CYANIDING GOLD AND SILVER ORES. A Practical Treatise on
the Cyanide Process ; its Application, Methods of Working, Design and Construction
of Plant, and Costs. By H. Forbes JuLLiN, Mining and Metallurgical Engineer ;
and Edgar Smart, A.M.I.C.E., Civil and Metallurgical Engineer.
' A complete story of cyaniding from its earliest beginnings to its latest developments.
* re been ' ' .... — .. - .
Practice and theory have been made to go hand in hand. The authors )peak with
authority."— CA«»mVj<rf Netcs.
" This very complete woric brings together in a single volume a mass of data which
should be invaluable to the Engineer."— Pa^e'9 Magazine.
Third Enolish Edition, Thoroughly Revised and greatly Enlarged. Price 7s. 6d.
THE CYANIDE PROCESS OF GOLD EXTRACTION. A TextBook
for the Use of Metallurgists and Students at Schools of Mines, dc. With Plates
and Illustrations. By Jahes Park. F.O.S., M.Inst.M.M., Professor of Mining and
Director of the Otago University School of Mines ; late Director Thames School of
Mines, and Geological Surveyor and Mining Geologist to the Government of New
Zealand.
" Deserves to be ranked as amongst the best of existing treatises on this subject."
—Mining Journal.
Extra Crown 8vo. With Folding Plate and 52 Illustrations. 3s. 6d. net.
LECTURES ON IRON-FOUNDING. By Thomas Turner, M.Sc.
A.R.S.M., F.I.C, Professor of Metallurgy in the University of Birmingham.
Third Edition, Revised. Volume I. 7s. 6d. net.
THE CHEMISTRY OF GAS MANUFACTURE. A Hand-Book on
the Production, Purification, and Testing of Illuminating Gas, and the Assay of the
Bye-Products of Gas Manufacture. By W. J. Atkinson Butterfisld, M.A., F.I.C,
F.C.S., formerly Head Chemist, Gas Works, Beckton, London, E.
Volume II. f completing the Work, Heady Immediately.
** The BEST WORK of its kind which we have ever had the pleasure of levlewing."-
Journal of Got Lighting.
LONDON: CHARLES GRIFFIN & CO., LTD., EXETER STREET, STRAND.
Digitized by V^jOOQ IC
ELECTRIC SMELTING
AND EEFINING:
THE EXTRACTION AND TREATMENT OF METALS
BY MEANS OF THE ELEGTRIG GURRENT.
BY
Dr. W^. B O R C H E R S.
Translatid fbom the Xuibd German Edition, with Additions, by
WALTER G. MCMILLAN, F.I.C., F.C.S.,
AUTHOR OF A ** TREATISE ON ELKCTRO-METALLUROY."
SECOND ENGLISH EDITION.
"^itb 4 |>[ate0 and flumeroua ^lluattationa in tbe tTest.
LONDON:
-CHARLES GRIFFIN AND COMPANY', LIMITED ;
EXETER STREET, STRAND.
1904.
[All Rights Reserved.]
Digitized by VjOOQ IC
Digitized by VjOOQ IC
(3 ^ ^ U\J
DhC 1 1904
PREFACE TO THE SECOND EDITION.
In this revised Edition of Dr. Borchers' well-known manual,
translated from advance proofs of the Third German
Edition, the progress made in Electro-Technology since
1897 has received full attention and has necessitated
re-writing a large portion of the volume. Every year
the task of supplying a comprehensive general account of
Electro-Metallurgical Practice becomes increasingly diflScult,
partly because of the great advances made by this branch
of metallurgy, and partly because of the secretiveness
imposed on manufacturers by the keenness of commercial
<M3mpetition. Being no longer engaged in industrial work,
the Author feels greatly indebted to the Metallurgists who
have favoured him with the results of their experience in
electro-metallurgical works. Moreover, in many cases
extensive plant is not required for investigating electro-
metallurgical processes on a large scale, and useful
information has been obtained in the Author's laboratory
at Aachen, where, by the liberality of the Prussian Minister
of Education, he has at his disposal 90 horse-power for
electro-metallurgical research.
Since the publication of the previous edition, numerous
authoritative works dealing with the newer theories of
Electro-Chemistry have been published; and the Author
has, consequently, omitted the introductory sketch of this
subject previously given.
Digitized by VjOOQ IC
VI PREFACE TO THE SECOND EDITION.
While this volume was passing through the press, the
untimely death of the Translator, Mr. Walter G. McMillan,
the accomplished Secretary of the Institution of Electrical
EDgineers, deprived it of the advantage of his final revi-
sion. Mr. MMillan combined exceptional knowledge of
electricity and metallurgy with conspicuous literary skill.
The Publishers feel confident, therefore, that the original
notes inserted by him here and there will be appreciated
by English readers as a valuable addition to the work,
and that the translation will be found to be an example
of the scrupulous care and accuracy that characterised all
his professional work.
The Publishers also desire to express their thanks to
Mr. Bennett H. Brough for kindly reading the final sheets
for press.
London, Jviy, 1904.
Digitized by VjOOQ IC
PREFACE TO THE FIRST EDITION.
In bringing before English readers an English version of
a manual so well known as that of Dr. Borchers — on a
subject which has so rapidly attained to the prominent
position now held by Electro -Metallurgy — but little by
way of introduction is needed.
The Author, in his preface to the Second German Edition
(that of 1896, from which the present volume has been
translated), refers to his twelve years' practical work in
chemical and metallurgical industries, and states that in
preparinor the first edition he had hoped to be of service
to his fellow-labourers by laying before them the results
of the experience with electro - metallurgical processes
which he had gained in conducting experimental tests on
a scale sufficiently large to enable him to form a trust-
worthy opinion as to their practical value. During the
four years which elapsed before the appearance of the
Second Edition, electro - technology had made enormous
progress; and at the end of that time the Author found
himself at liberty to publish much information which pre-
viously, owing to personal considerations and business
obligations, he had been prevented from giving to the
world. He, therefore, re-wrote and greatly extended the
work in the edition from which this version is taken.
A short sketch only is given of the newer electro-chemical
theories, and the student is referred for fuller and more
elaborate explanations to the works of Ostwald on General
Chemistry and Electro-Chemistry, and to that of Nemst*
• Nernst's TheoreticcU Chemistry from the Standpoint of Avogadro'a RuU
uud TliermodynamicB, Translated by C. S. Palmer.
Digitized by VjOOQ IC
Till PREFACE.
on Theoretical Chemistry. Those who are engaged in
scientific or technical research are recommended to consult
Ostwald's Hand- und Hiilfabuch zur Auafuhrung physiko-
cheniischen Meaaungen* and Oettel's Anleitung zu Elektro-
chemiacheji Versuchen.
In the present volume, all those metals in the extrac-
tion and working of which the electric current has
found any application are treated of; but electrolytic
analysis, electro-plating, and electro-typing have not been
touched upon. From the mass of material accumulated
in journal and patent-ofBce literature, only those papers
or processes which are capable of practical application
have, usually, been selected for reference. A short
survey of the purely metallurgical methods of treating
the metals has been added to each chapter, so that the
reader may be in a position to compare such methods with
electro-metallurgical processes, and to see how the two
may be used in conjunction.
During the few months that have passed since the
publication of the German Edition, progress has been
made in many directions. The Tran.slator has, therefore,
ventured to add notes here and there in the hope of
briuging the work more nearly up to date. He has
also added accounts of the practical working of certain
processes which have been developed since the German
work was placed before the public. Extra references to
English patents and the British equivalents of the
metrical measurements quoted in the text have also
been given. All these additions, whether in the text
or in footnotes, are enclosed within square brackets, [ ],
so that the reader may distinguish between the original
and the supplementary matter. Tables for the conversion
of certain Continental into the corresponding British units
have also been added in an appendix. Whenever possible,
quotations from the French have been translated direct
* Ostwald's MantujU of Physico - Chemical Measurements. Translated
by J. Walker.
Digitized by VjOOQ IC
PREFACE. IX
from the French text, instead of from the German version,
and those from English sources have been reprinted ver^
baiim from the original.
In arranging the additional matter, the Translator has
kept in view the industrial aspect of the question, and
has introduced one or two short supplementary notes
referring to actual applications of processes not otherwise
referred to in the text. It is believed, therefore, that
under each of the various metals, separately, enough has
been stated to give a good general idea of the present
industrial position of that branch of the subject.
The practical applications of Electro-Metallurgy are
advancing with phenomenal rapidity, especially on the
Continent and in America, and there is little doubt that,
if England is to hold her own in Metallurgy, she also must
give more and more attention to the electro-metallurgical
smelting and refining processes, which are cheapening and
simplifying the work of producing high-class products
abroad.
In conclusion, the translator's thanks are due to Mr. A.
E. Hunt and the Institution of Civil Engineers, to the
Electrician, and to Industries and Iron, for the use of
the blocks employed in Figures 82a, 101a, and 50a and
50b respectively.
WALTER G. MCMILLAN.
MfiTALLDBOIOAL DEPARTMENT,
Mason College,
BlRMiKOUAM, June 8, 1897*
Digitized by VjOOQ IC
Digitized by VjOOQ IC
CONTENTS.
FABT I.— ALKALIS AND ALKALINE -EABTH
METALS.
Chapter I.— Magnesium.
Occurrence in Nature,
Properties of the Metal, .
Early Experiments in Reduction,
Bnnsen's Electrolytic Reduction
Process, ....
Matthiessen's Double Chloride
Process,
Berthaut's Process, .
Small-scale Apparatus,
Fischer's Apparatus,
GraetzePs ,,
i!
3,
Napier's Electrolytic Vessel, .
Borchers' Magnesium Process,
Preliminary Preparation of the
Charge, ....
The Roauction Process, .
Melting of Electro -deposited
Magnesium, ....
Electro -thermic Reduction of
Magnesium,.
Hilberg's Process, .
Applications of Magnesium, .
PAGE
12
13
16
18
19
19
20
20
Occurrence in Nature,
Properties,
Extraction,
Troost 8 Process,
Hiller's Apparatus, .
Chapter II.— Lithium.
21 Grabau's Apparatus,
. 21 : Guntz's Process, .
21 1 Treatment of Lithium Minerals,
22 1 Kahlenberg's Process,
23 Uses of Lithium,
25
26
26
27
28
Chapter IIL—Sodium.
Occurrence in Nature,
Properties,
Reduction Processes,
Electrolysis, .
Davy's Experiments, .
Watt's Process, .
JablochkofTs Process, .
Hoepfner's ,,
Rocers*
Omholt's Apparatus,
Fischer's Process,
Homung and Kasemeyer's
Plant,
Grabaa's Apparatus, .
Borchers' „
Grabau's later Apparatus,
29,
29 I
30 '
30
30
31
33
34
33
36
40
41
42
45
47
Borchers' Sodium- Ex traction
Plant 47
Danckwardt's Apparatus, . 49
Electrolysis of Fused Salt, . 52
Mutual Action of Sodium and
Chlorine, .... 52
Fischer's Modifications, 53
Formation of Sodium Alloys, 58
Borchers' Plant for Sodium
Alloys, .... 60
Castner's Process for Electro-
lysis of Caustic Soda, . 64
Becker's Apparatus, . 65
Rathenau and Suther's Process, 67
Darling's Process, 68
Uses of Sodium, ... 70
Digitized by VjOOQ IC
CONTENTS.
Chapter IV.— Potassium.
Occurrence in Nature,
PAGE
71
Extraction Processes,
PAGE
71
Properties, ....
71,
Uses,
72
Chapter V.— Calcium,
Strontium, Barium.
Occurrence in Nature, .
73
Borchers and Stockem^s Cal-
Properties of the Group, .
73 1
cium Process,
83
Extraction,
74'
Production of Strontium,
85
Matthiessen's Experiments,
75 1
„ of Barium, .
86
Extraction of Calcium, .
76'
Reduction of the Alkali-Earths
,, of Strontium,
78 1
by Carbon, ....
86
„ on a large Scale,
79
Uses of the Alkali- Earth Metals,
87
Chapt(
Br VI.-
-Beryllium.
Occurrence in Nature,
87 1
Liebmann's Process, .
89
Properties,
87 1
Warren's ,,
90
Extraction,
. 88 1
Reduction of Beryllium Oxide,
90
Electrolysis, .
. 88 1
Electro-thermic Method,
90
Lebeau's Process, .
89
Lebeau's Experiments, .
91
PABT II.— THE EABTH METALS.
Chapter I.— Aluminium.
Occurrence in Nature, . . 93
Properties, .... 93
Precipitation Processes of Ex-
traction, ....
The Reduction Processes,
The Cowles Process,
Theory, ....
Thermic Reduction of Alumina
by Carbon, . . - .
The Electrolytic Methods of
Reducing Aluminium,
Electrolysis of Aqueous Solu-
tions, . . . .104
Early Experiments, .108
Deville's Process,. . . 109
Practical Obstacles to the Elec ■
trolysis of the Chlorides, .
The Teaching of Deville*s and
Bunsen's Experiments,
Impracticable Processes,
Successful Reduction Processes, 1 19
Decomposition of Aluminium
Compounds by the Electric
Arc,
Johnson's Process,
The Siemens Electric furnace,
The Kleiner- Pi ertz Electric
Furnace, . . .123
95
96
98
101
103
104
112
112
113
119
119
120
TheGrabau Electric Furnace,
The Gerard - Lescuyer Fur
nace,
The Willson Furnace, .
Electrolysis of Fuseid Alu
minium Compounds, .
The H^roult Process, .
The Hall Process,
The Minet-Bernard Process,
The Neuhausen-Froges Pro
cess, ....
Borchers' Aluminium Extrac
tion Apparatus,
Bradley's Process,
Kiliani Furnace, .
Grabau's Process,
The J. B. Hall Process,
Electrolysis of Aluminium
Sulphides,
Preliminary Treatment of
Bauxite and Cryolite,
Industrial Electrolysis of
Aluminium,
Conditions to be Observed in the
Extraction of Aluminium,
Aluminium Works, .
Uses, ....
Price
125
125
125
126
127
131
141
145
145
148
151
153
153
154
155
157
160
161
162
165
Digitized by VjOOQ IC
CONTENTS.
Chapter II.— Cerium, Lanthanum, Ppaseodidymium,
Neodidymium.
Occurrenceof theCerite Metals,
Properties of Cerium,
Electrolytic Deposition of the
Cerite Metals,
Stockem*s Experiments in the
Reduction of Cerium, .
PAGE ,
166 j Muthmann, Hofer, & Weiss's
167 Experiments,
I Reduction Process, .
167 I Reduction of Cerium Alloys,
, Extraction of Cerium,
170 ' Uses of the Cerite Metals,
171
180
181
182
182
PART III.— THE HEAVY METALS.
Chapter L— Copper.
Occurrence in Nature, . 184 i
Sketch of Metallurgy, . . 185
Properties, . . • . 186
Electrolytic Methods, . . 187
Electrolytic Refining, . .187
Early attempts in the Electro- |
Deposition of Copper, . 187
Patera's Process, . .188
The Elkinffton Patents, . 188
Theory of Electrolytic Copper
Refining, .... 192
Kiliani's Experiments, . 193
Wohlwiirs Experiments, . 199
Behaviour of Chlorine in the
Electrolyte, . . .203
Crystalline Growths on
Cathodes.. ... 205
Porster and SeideFs Experi-
ments, . .207
Von Hubl's Experiments, . 208
Efiect of Temperature of
Bath on Copper, . 209
Use of Alcohol in Copper
Solutions, . .210
Development of the Electro-
lytic Copper-refining In-
dustry 210
General arrangement of Plant, 211
Parallel System, .211
Arrangement of Electrodes in
Baths, . . .211
Casting of Anode-plates, . 213
Anode-casting Machines, 214
Cathode-plates, . .218
Depositing Vats, . .218
Borchers' Improvements, . 221
Details of the Siemens-
Borchers' Installation, . 224
Reduction in Cost, 225
Schneider and Szontagh
System, .... 227
The Thofehrn Process .
Thofehm*8 New Process,
Raritan Copper Works Plant,
Connections of Electrodes, .
Series System,
The Stalmann System,
The Hayden, Smith, and
Randolf Processes, .
The Multiple and the Series
Systems, ....
Treatment of Impure Electro-
lytes, ....
Process at Perth Amboy
Works, ....
Process at Raritan Works, .
Smith Process,
Practice at the Chicago
Copper Refinery,
Treatment of Anode Slimes, .
Treatment at Works of Bal-
timore Electric Refining
Co.,
Treatment at Raritan Copper
Works, ....
Industrial Conditions in Elec-
trolytic Copper Refining, .
' Treatment of Sulphides and
Ores, ....
The Marchess Process,
The Stolbere Installation,
Use of Depolarisers, .
Body's Process, .
The Siemens- Halske Matte-
Refining Procesp.
Modified Siemens • Halske
Process, ....
The Hoepfner Process,
The Schwarzenberg Experi-
ments, ....
The Coehn Process,
Applications of Copper,
228
233
234
235
236
236
238
239
240
241
241
241
242
243
244
244
245
246
247
254
257
259
260
263
266
270
271
272
Digitized by VjOOQ IC
XIV
CONTEXTS.
Chapter IL— Nickel.
PAGE
Occurrence, .... 273
Propei-ties of Nickel, .274
Extraction, .... 275
Andres Process, . . 275
Classen's Experiments, 277
Farmer's Process, . .277
The Basse-Selve Process, . 279
The Strap Processes, . . 280
The Rickets Process, . . 280
The Hoepfner Process, . 280
The Milnzing Process, . 282
Deposition of Pure Nickel
and Cobalt, . . .282
The Heibling Process, . . 284
The Le Verrier Process, . 284
PAGE
The Kugel Process, . . 284
The Frasch Process, . . 284
The Vortmann Process for
Separating Cobalt, . . 285
Ulke's Experiments, . . 286
Forster's Experiments, 287
Separation of Nickel -Copper
Alloys, .... 288
Review of Processes, . . 290
Treatment of Concentrated
Mattes, . . . .291
Borchers* Process for Separa-
tion of Iron and Nickel, . 293
Utilisation of Anode Reactions, 294
Summary of Processes, . . 296
Chapter III.— Silver.
OccuiTence in Nature,
Properties, ....
Extraction by solution of Silver
in another metal, .
Separation by Processes of
Chemical Solution, .
The Ziervogel Process,
The Augustin, Patera, and
Kiss Processes,
Dietzel's Process,
Solution of the Constituents
other than Silver,
300 Production of Fine Silver, . 311
300 Wohlwill's Process, . .311
Moebius Process, . - .317
301 The Pennsylvania Lead Co.'s
L Plant 322
303 I Later form of Moebius Ap-
303 paratus, .... 323
The Perth Amboy Installation, 324
303 ' Moebius and Nebel's Ap-
303 paratus, .... 326
The House- Symon Process, . 327
309 Balbach's Apparatus, . . 327
Chapter IV.— Gold.
Occurrence in Nature, 329
Properties, .... 329
Mechanical Treatment, . , 330
Solution in other Metals, . 331
Treatment with Lead, .331
Principles of Amalgamation, . 331
Electro- Amalgamation, . . 332
MoUo/s Process, . . . 332
Breakell and Haycraft Pro- i
cess, 333 j
Barker's Process, . . . 333 |
Hannay's ,, . . 334
Bailey ^s „ . . .336
Rae's „ .336
Pelatan and Clerici's Process, 336
Motz's Process, . . .336
Jory's ,, . . .338
Leachingand Precipitation Pro-
cesses, .... 338
Chlorination, . . 338
Cassel Process, . . 338
Body's ,, . . .339
Stolp's „ . . .339
Cyanide Process, . . . 340
Theory. .... 340
Practical Development, . 341
Siemens- Halske riant, . 351
Comparison of Electrolytic
and Chemical Precipitation
Processes,. . . . 354
Refining, 35S
Wohlwill's Process, . . 359
Uses, 382
Digitized by VjOOQ IC
(
CONTENTS.
XV
Chapter V.
—Zinc and Cadmium.
PAGE 1
PAGE
Zinc,
382 i
Leaching and Precipitation,
408
Occurrence in Nature, .
382
L^trange Process, .
Other Patented Processes,
408
Properties, ....
383
409
Extraction, ....
384
Coehn's Process,
412
The Roasting and Reduction
Cowper-Coles' Process, .
412
Process,
384
Eschellmann's Process, .
412
The Cowles Zinc Process, .
384
Treatment of Highly Sili-
cesses, ....
413
ceous Ores, .
385
Luckow's Process, .
413
Treatment of Blende-bear-
Ashcroft's Process, .
414
ing Pyrites, .
387
The Phcenix Process,
415
Leaching and Precipitation
Methods,
Hoepfner's Process, .
418
388
Electrolytic Deposition of
Luckow Process,
388
Zinc from Fused Salts, .
441
Experiments by Dorsema-
gen and Borcners,
Pfleger's Process, .
442
389
Uses, ....
445
Nothemann's Process,
395
Cadmium-
Refining of Alloys,
Mylius and iromm's Ex-
395
Properties, ....
446
Extraction, ....
446
periments, .
398
Electrolysis,
446
The Work of Forster and
Uses,
447
Giinther,
403
.—Mercury.
Chapter VI
Properties and Occurrences, .
447
lUses,
448
Extraction, ....
447
1
Chapter VII.~Tin,
Properties and Occurrence, . 449 Electro-chemical Treatment
Extraction, .... 449
Calcining and Reducing Pro-
cess, 450
Leaching and Precipitation, . 450
of Tin-plate, Scrap, and
Tin Residues, ... 450
Refining, 459
460
Chapter VIIL— Lead.
Properties and Occurrence,
Extraction, ....
Smelting of the Lead Ore,
Roastmg Process,
Roasting and Reduction Pro-
cess, ....
Precipitation Process, .
460 I Leaching and Precipitation, . 462
461 ; Lyte's Process, . . .464
461 I Salom's Process, ... 464
461 Manufacture of Pure Lead, . 466
The Keith Process, . . 466
462 The Tommasi Process, . . 469
462 Applications, . . . .470
PropertiM and Occurrence,
Extraction,
Liquation, .
The Reduction Process,
The Precipitation Process,
Chapter IX.— Bismuth.
Oxidation Process, . 473
The Sulphide Process, • 473
Electrolytic Process, . • 473
Applications, . . . • 474
470
471
471
472
472
Digitized by VjOOQ IC
CONTENTS.
Chapter X.— Antimony and Vanadium.
Antimony, . . . .
Properties, . . . .
Extraction, . . . .
The Reduction Process, .
The Precipitation Process,
Electrolytic Processes,
Borchers' Experiments,
Practical Working,
Conditions Suitable to
Electrolytic Refining,
Sanderson's Process,
PAGE
474
474
475
476
475
476
476
479
484
486
PAGE
Kopp's Process, . 486
Siemens-Halsko Process, 487
Treatment, .... 487
Applications, 487
Vanadium, .... 488
Occurrence and Properties, . 488
Treatment of Ore, . 488
Reduction Process, 489
Electrolysis, . 4S9
Uses 490
Chapter XL— Chromium, Molybdenum, Tungsten, Uranium,
Manganese.
Chromium,
. 490
Reduction of Oxide, .
603
Occurrence in Nature,
. 490
Uses,
604
Properties, .
. 490
Tungsten, ....
604
The Reduction Process,
. 491
Properties
604
Moissan's Furnace,
. 492
Treatment of Ores,
604
Chaplet's „
. 492
Reduction Process,
605
Refining,
. 493
Uranium, ....
506
Heibling's Process,
. 494
Properties, ....
606
Aschermann's Process,
. 495
Precipitation Process, .
506
The Precipitation Process,
Goldschmidt ,,
. 495
Reduction ,,
506
. 495
Electrolytic Methods, .
507
Electrolysis,
. 495
Uses
607
Dissolved or Fused Chro
-
Manganese, ....
507
raium Compounds, with
Occurrence and Properties, .
607
Insoluble Anodes,
. 495
Reduction, ....
608
Bunsen's Experiments,
. 496
The Precipitation Process, .
608
Borchers' „
. 496
Electrolysis,
609
The Placet-Bonnet Process. 498
Dissolved or Fused Man-
MoUer and Street's ,,
499
ganese Compounds, with
Insoluble Anodes,
Feree's Experiments,
. 501
609
Glaser's ,,
. 601
Fused Manganese Com-
Krupp's Process,
. 602
pounds, with Soluble
Uses, ....
. 602
Anodes,
609
Molybdenum, .
. 502
H^roult's Process, .
510
Occurrence and Properties,
. 602
Applications,
611
Production, .
. 603
Chapter :ij
:iL-Iron.
Occurrence in Nature, .
611
Hdroult's Process,
615
Properties,
611
Harmet's ,, . .
515
Preparation of the Raw Ma
Keller's „ . .
617
terials.
613
The Production of Malleable
Magnetic Separation,
613
Iron, ....
619
Production of Pig Iron, .
The Stassano Pig-iron Pro
613
Siemens' Furnace,
619
De Laval's Smelting Furnace,
619
cess, ....
614
The H^roult Process, .
621
Digitized by LjOOQ IC
CONTENTS. XVU
Chapter XIII.— Hetallie Compounds : Carbides and Sllieides.
Carbides, .
PAQB
. 531
The Electric Furnace, .
PAGE
. 635
Historical, .
. 531
Consumption of Power,
Alkali Carbides, .
. 542
Raw Materials, .
. 534
. 544
Calcium Oxide, .
. 534 1 Other Carbides, .
. 544
Carbon, .
. 535 i Silicides, ....
. 546
ADDENDA.
Table I.— Value of Equal Current Volumes 547
,, II. — ^Thermometer Scales, 548
Index, 549
Digitized by VjOOQ IC
Digitized by VjOOQ IC
LIST OF ILLUSTRATIONS.
Plate L Copper refinery, SiemeoB' system,
to face p.
n. „ Siemens-Borchers-Brothers' system,
m. H^roult fumaoe in operation,
IV. Keller fumaoe, .....
221
224
522
540
1. Bonsen's electrolyte cell, .
2. „ carbon electrode,
3. Gomp-Besanez apparatus,
4. Fischer's apparatus, 1884,
5. 6. Graetzel's apparatus,
7, 8« „ ,,
9. Borchers* flanged crucible for electrolytic reduction,
10. , , experimental furnace for the electrolysis of fused salts,
11. ,, fumaoe for use with currents of 100 amperes, .
12. „ fumaoe for the electrolysis of magnesium compounds,
13. Troost's apparatus for the electrolysis of fused lithium chloride,
14. Killer's apparatus for the extraction of lithium,
15. 16. Grabau's electrode cell for use with light metals,
17. Watt's apparatus for the extraction of sodium,
18. JablochkofiTs „ „ „
19. 20. Rogers' „
21, 22, 23. Omholt's sodium plant,
24. Omholt's modified sodium plant, .
25. Fischer's apparatus,
26. 27. Homung and Kasemeyer's apparatus,
28. Grabau's apparatus for the reduction of alkali metals,
29. Borchers' apparatus for reducing the alkali metals,
30. Grabau's modified cathode cell,
31. 32. Borchers' sodium-extraction apparatus,
33-37. Danckwardt's „ „
38. Diagram illustrating behaviour of sodium and chlorine in elec
trolysis of fused sodium chloride,
30, 40. Hood to surround cathode in sodium-chloride electrolysis,
41. Fischer's sodium-chloride electrolysis apparatus,
42. „ modified sodium-chloride electrolytic apparatus,
43. Special form of cathode, .
PAGI
3
3
5
6
8
9
12
13
14
15
22
24
25
36
40
41
43
45
47
48
50
53
53
54
55
56
Digitized by VjOOQ IC
XX LIST OF ILLUSTRATIONS.
Fia. PAGE
44. Arrangement of cathode, ...... 57
45, 46. Borchers' electric furnace for experimental work, 58
47. Vautin's apparatus, ...... 60
48. Borchers* apparatus for the production of sodium alloys, . 61
49. „ improved apparatus for the production of sodium alloys, 62
50. 51. Apparatus for the production of lead-sodium alloys, 63
52. Castner's electrolytic sodium-extraction plant, . . . 64
53-55. Becker's modification of same, ..... 65
56. Darling's apparatus for the extraction of sodium from sodium
nitrate, ........ 69
57* Borchers' experimental apparatus for alkaline -earth metal re-
duction, ........ 80
58, 59. Borchers and Stockem's calcium -extraction furnace, . . 82
60, 61. ,, „ modified calcium -extraction furnace, 83
62. ,, „ strontium-extraction furnace, . 85
63. Cowles' plant for manufacture of aluminium alloys, . . 99
64. The Cowles' furnace (longitudinal section), . . 100
65. „ „ (cross-section), .... 100
66. Arrangement to show direct reduction of alumina by carbon, . 103
67. Deville's apparatus for reducing aluminium, . .111
68. Graetzel's aluminium electrolysing vessel, . .115
69. Grabau's cooled cell apparatus, . . . .117
70. Cowles' ore-smelting furnace, ..... 120
71. Pichou's smelting furnace, .121
72. Siemens' electric furnace, . . . .121
73. „ modified electric furnace, .... 123
74. 75. The Kleiner-Fiertz furnace, . .124
76. Gerard-Lescuyer's electric furnace, , .125
77. The Willson furnace, . . . .126
78. 79. The H§roult aluminium furnace in longitudinal section and
in plan, ........ 128
80. H^roult aluminium furnace for small installations, . 130
81. The Hall furnace, U.S.A. patent, No. 400,766, . .132
82. Modified apparatus. Hall patent, 400,766, . . .133
83. 84, 85. The Hall furnace, American patent. No. 400,664, 134
86. „ „ „ 400,665, 135
87. „ „ „ 400,666, 137
88. The Pitteburg Co. 's Hall furnace, . . .138
89. 90. Bernard (Minet) aluminium furnace, . . .143
91. Aluminium reduction. Apparatus used at Neuhausen, . .145
92. Borchers' aluminium extraction apparatus, . 146
93. Modified Borchers' aluminium furnace, .... 147
94. 95. Bradley's aluminium smelting furnace, . . 150
96. Kiliani's furnace, . . . . .151
97-101. Schindlers aluminium furnace, .... 152
102. Section of Borchers' furnace, ..... 172
Digitized by LjOOQ IC
LIST OP ILLUSTRATIONS.
XXI
103. Section of Mnthmann, Hofer, and Weiss's fnmace,
104, 105, 106. Furnace for extraction of cerium metals,
107. Borchers* furnace for cerite metals,
108, 109. Stockem'B furnace for production of cerium alloys,
110. Diagram illustrating irregular solution of copper anodes,
111-117. Typical anodes and cathodes,
118, 119. Mould used by Morrow for casting anodes,
120, 121. Hixon & Dyblie's anode-casting machine,
122. Walker's „ „ „
*^'' it If y» >»
124. Section of copper-refining vat, showing anode (Siemens),
125. Cross-section of copper vat (Siemens), showing cathode,
126. Longitudinal section of copper vat (Siemens),
127. 128, 129. Plate I., to/aa
130. Scheme of electrical connections in copper vats (Siemens),
131. Siemens-Borchers' copper vats (longitudinal section), .
132. „ „ (cross-section), .
133, 134, 136. Plate IL, to face
1 36. The electrical connections of Thofehm's apparatus,
136a. The Anaconda Thofehm installation, .
137, 138, 139. Stalmann's electrode connections,
1 40. Stalmann's copper vat (cross-section),
141. Scheme of electrical connections in Stalmann's vat,
142. „ „ „ Hayden*B ,,
143-146. Moulds for casting anodes of matte at Casarza,
147, 148. Connection of anode-strip with main conductor at Casarza,
149. Scheme of electrical connections at Casarza,
150, 151. Method of attaching the lead linings to the vats (Casarza),
152. Section of vats, showing method of circulating the electrolyte
(Casarza), .......
153. Plan of vats, showing method of circulating the electrolyte
(Casarza), ....
154. Plan of the Casarza installation, .
155. Cross-section of the Casarza depositing-room,
156. Interior of the Casarza depositing-room,
157. General scheme of the Marchese process,
158. 159. Body's apparatus,
160. Siemens k Halske electrolytic cell for treating copper ore,
161. Arrangement of vats in the Siemens-Halske process,
162. 163, 164. Siemens & Halske electrolyte cell (1889),
165. New form of anode (Siemens-Halske), .
166. Newer form of tank ( „ ), .
167. Coehn's single-compartment electrolyte cell,
168. 169. Farmer's nickel depositing plant, .
170-173. Apparatus for the preparation of lead chlorate at the anode
and nickel at the cathode, . . . • •
PAaz
172
176
179
180
201
212
213
215
216
217
218
219
220
221
221
222
223
224
228
231
236
237
237
238
248
249
249
250
250
250
251
251
252
263
269
261
262
263
265
265
271
278
295
Digitized by VjOOQ IC
xxu
LIST OF ILLUSTRATIONS.
no.
174. Dietzers alloy-separating vat, .....
175. The Moebitis apparatus, showing anode compartment in half
section, ......
176. The Moebius apparatus—section showing cathode,
177. „ „ section through A R C E (Fig. 178),
178. „ ., plan, ....
179. „ ,, scheme of electrical connections,
180. 181. The Moebius process — ^latest form of apparatus, .
182. The Moebius anode connection (new form),
183. Moebius and Nebel's modified apparatus,
184. Balbach's electrol3'8is tank,
185. „ n ft
186. The MoUoy amalgamating pan, .
187. 188. Hannay's electro-amalgamating vats,
189, 190. Motz's electrolytic sluice,
191. Stolp's gold-extraction apparatus,
192. „ modified apparatus,
193. 194, 195. Construction of wooden separating vats,
196, 197. „ iron „ „
198, 199. Section and plan of vat, showing Butters and Mein's dis-
tributors, ....
200-203. Bottom discharge doors for tanks, .
204-210. Details of side discharge doors for tanks,
211-213. Illustration of electrolytic tanks for use in Siemens-Halske
cyanide process, .....
214. Cowles' electrically-heated zinc retort, .
215. Dorsemagen's zinc and silicon carbide furnace, .
216. Experimental chlorination barrel used by Dorsemagen, .
217. Chlorination cylinder to be used in practice,
218. Connections with subsidiary electrodes in zinc deposition,
219-222. Apparatus for electrolysis of zinc solutions, .
223. Hoepfner's electrolytic tank for zinc chloride solution, .
224,225. „
226-228. Keith's plant, .
229,230. „ lead-refining vat,
231. Antimony extraction plant (section),
232. „ „ „ (plan),
233. ,, electrolytic tank (section),
234. „ „ „ (plan),
235. Borchers' electric furnace,
236. Moissan's ,,
237. Chaplet's
238. „ modified electric furnace,
239. Harmet's electrical smelting plant,
240. Keller's arrangement,
241. 242, 243. De Laval's electric furnace,
PAGE
304
Digitized by LjOOQ IC
LIST OF ILLUSTRATIONS.
TXlll
FIO.
244. Hdroult fiiniace, ....
245. ,, „ io operation (Plate UL),
246. 247. Steel furnace at Gysisge,
248, 249. Stassano furnace,
250. Early fl^roult ,.
251. Clarke electric ,,
252. Tap
253. Keller famace (Plate IV.),
254. ft f9 • •
255. Chart ahowiog oonsamption of power.
PAGB
521
facing 522
. 524
526
536
538
539
faciag 540
541
543
Digitized by VjOOQ IC
Digitized by VjOOQ IC
ELECTRIC SMELTING AND REFINING:
PART I.— ALKALI- AND ALKALINE-
EARTH METALS.
In ihe following pages, the alkali- and alkaline-earth metals
are claissified according to their behaviour during the process of
electro-chemical separation, because, having regaiS to their pro-
duction technically, such a treatment of the subject is more
practical than one in which the general properties of the metals
form the bfusis for classification.
FIRST GROUP.— MAGNESIUM, LITHIUM.
CHAPTER I.
MAGNESIUM.
Ooourrence in Nature. — ^As might be expected from the
properties of the metal, magnesium is found in nature only
in the form of salts. It occurs as a haloid salt in camailite
(MgCLj.Ka. 6H2O) and kainite (MgCl.^ . MgSO^ . K2SO4 . 6H2O) ;
as sulphate in laCBerite (Mg80..'H^0) ; as carbonate in mag-
neaite (MgCOo) and dolomite ^MgCOg . nCaCOg) ; and as silicate,
always in combination with other silicates, in asbestos, steatite,
serpentine, talc, meerschaum, and many other minerals.
Properties of the Metal. — Magnesium (Mg" ; atomic weight
= 24 j specific gravity = 1 -75) is a white lustrous metal with a
fibrous crystalline structure ; it is so far malleable that it may
be rolled into wire or ribbon, but by reason of its relatively low
toughness it can readily be reduced by filing to a fine powder, a
property which greatly favours the use of the metal in pyro-
Digitized by LjOOQ IC
2 BLECTRIC SMELTING AND REFINING,
techny. It melts at a temperature of 500* to 600® C, and boils
at temperatures exceeding 1,100"* C. Large and thick pieces of
the metal are but little altered by exposure to the air, although
they may become superficially oxidised, especially in the presence
of moisture; they may even be melted in an open crucible
without risk of burning. In the form of a fine powder or thin
plate, however, magnesium oxidises very readily, and at a higher
temperature it burns with a brilliant light, which is particularly
rich in the chemically active rays. In the presence of only a
restricted quantity of air, magnesium nitride is always formed
as well as oxide. The property possessed by magnesium of
combining with nitrogen, at a moderate red heat, might be of
great practical value in the synthesis of nitrogen compounds, if
a sufficiently cheap method of producing the metal should be
discovered.
The powdered metal also, when once moistened, cannot be dried
without undergoing complete oxidation. At the ordinary tem-
perature it readily decomposes water which contains even a
small proportion of dissolved salts, but it has less effect on
pure water. Finely-divided magnesium burns freely in super-
heated steam, in sulphur, and in the halogens. Magnesium
dissolves readily in most acids and salts; in the latter case it
either displaces the metal which the salts contained previously,
or it forms basic compounds with the salts themselves. Its
electro-positive solution-pressure is so great that the metal is
able to liberate, not ordy other metals, but even metalloids,
from their compounds, so that, for example, carbon monoxide,
carbon dioxide, silica, and boron trioxide are reduced by it,
with separation of carbon, silicon, and boron respectively.*
On account of the strong reducing-action of magnesium, par-
ticularly on oxygen compounds, special care must be devoted
to the preparation of the substances to be used in the extraction
of the metal. Just as the oxygen compounds of magnesium are
avoided in the selection of the materials from which the metal
is to be extracted, it is necessary that all accidental traces of
oxygen compounds, such as water, sulphate, <kc., should be most
completely removed from the compounds (for example, haloid
salts) to be used. The double magnesium and alkali-metal
chlorides are used for magnesium reduction in preference to
the haloid, and indeed to all other, salts of magnesium, not
only on account of their chemical and electrical properties, but
also by reason of the comparatively low cost of producing them.
It must be remembered that although the bulk of the water of
crystallisation of these salts is expelled with comparative ease,
small quantities are retained even at a bright red heat.
* See also CI. Winkler on " The Redaction of Oxygen Compounds by
Magnesium," BerichU der detU9e?ien chemische OestlUchafi, voL zziv.,
pp. 873 and 1966.
Digitized by VjOOQ IC
MAGNESIUM. 3
Magnesia, like all other oxides that were, up to the year 1891,
held to be unreducible by carbon, is capable of reduction through
the agency of electrically-heated carbon. This is readily demon-
strated with the aid of the small electric fum€k« described in
the first edition of Borchers' ElektrometaXlurgie ; * but no appre-
ciable quantity of the metal can thus be obtained, owing to the
action of magnesium on oxygen compounds already referred to.
Sarly Experiments in Redaction. — It appears, therefore,
unqu^tionable that Davy's experiments,! in which he reduced
white-hot magnesia by means of potassium vapour, did not yield
pure magnesium. This view is confirmed by the description
given by Davy of the properties of the metal obtained. Bussy,t
Bufr,§ and Liebig|| obtained pure metal because they adopted
Wohler's aluminium process, decomposing the chloride by means
of potassium.
Fig. 1. — Bunsen's Fig. 2. — Bunsen's
electrolj'^tic cell. carbon electrode.
Bunsen's Electrolytic Reduction Process. — Bunsen was
the first to recognise the possibility of decomposing fused mag-
nesium chloride by electrolysis, and to indicate the manner in
which other metallic chlorides or halogen compounds in the
fused and anhydrous condition might also be decomposed into
metal and halogen. In 1852 he published a notell on this
subject, of which the following is a translation : — " Fused mag-
nesiiun chloride is so easily decomposed by the current, that
several grammes of heavy metallic regulus may be obtained from
it in a short time with the aid of a few carbon-zinc elements.
"The magnesium chloride is best prepared by Liebig's well-
known method. For the decomposing cell it is convenient to use
a porcelain crucible (Fig. 1) about 3| ins. high and 2 ins. wide,
, divided into two compartments by a diaphragm, reaching from
* Published Leipzig, 1891 ; see p. 62.
t PhU, Trans,, London, 1808, p. 336.
XJourn, de Chim. midicale, 1849, vol. vi., p. 141.
I Pogg. Ann,, vol. xviii., p. 140.
II Pogg, Ann,f vol. xix., p. 137.
t Liibig'a Ann., 1852, vol. IxxxiL, p. 197.
Digitized by VjOOQ IC
4 ELECTRIC SMELTING AND REFINING.
the top half way to the bottom ; in one of these the Uberated
chlorine escapes upwards, and is thus prevented from obtaining
access to the magnesium deposited in the other. The diaphragm
may consist of a thin porcelain tile, broken to the required
shape with the aid of die notches of a key. The crucible is
covered with a lid made of ordinary fire tile, filed into shape,
and bored with two holes (Fig. 2) through which the poles are
passed. These poles are cut out of the material from which
the battery carbons are prepared, a process that presents no
difficulty, as such carbons may easily be bored, turned on the
lathe, filed, and even provided with a screw thread. The
carbon poles are fastened to the cover by means of the
wedges 5, g?, which are also of carbon, and which serve to grip
the platinum strips that are used to convey the current to
and from the apparatus. The saw-like notches in the nega-
tive pole retain the reduced metal which collects in them
in the form of regulus. In the absence of this arrangement,
the metal would float to the top of the relatively heavier liquid
and would there be at least partially burned. In conducting
the experiment the crucible, with the cover and the poles
attached to it, are first heated until they are red hot, they are
next filled to the brim with fused magnesium chloride, and the
electric circuit is then completed."
Bunsen quotes by way of example an experiment in which the
current from 10 carbon-zinc elements was allowed to operate
for 115 minutes. From the measurements that were made at
intervals of five minutes, he calculated the theoretical quantity
of reduced magnesium at 4*096 grammes, which would corre-
spond to a current of 4*7 amperes. Since no details are given
of the size and arrangement of the battery cells, it is impossible
to ascertain what electromotive force was used for the decom-
position in this experiment.
Matthiessen's Double -Chloride Process. — Matthiessen's
proposal* to substitute the double-chloride of magnesium and
potassium (camallite) for the single magnesium chloride, which
is with difficulty obtained anhydrous, is especially wolrthy of
notice, as it has been adopted at a later date for the production
of magnesium on a manufacturing scale. The carnallite, as it
occurs in nature, cannot, however, be so used direct, by reason
of the sulphates and other impurities that it contains, which
interfere with the dehydration, fusion, and electrolysis of the
salt.
Berthaut's Process. — Berthaut's patented process f consisted
in the use of a soluble anode made by compressing a mixture of
carbon and magnesia, and employing it in a bath of magnesium
chloride. This was an imitation of Deville's proposal J to use
• Journ. Chem. Soc., vol. viii., p. 107.
t Eng. Patent 4,087, of 1879. % See Aluminium,
Digitized by V^jOOQ IC
MAGNESIUM.
an anode composed of aluminium and carbon for the reduction of
aluminium. Electrodes of this description, however, are liable
to fall to pieces owing to the extraction of the oxide, and thus
to contaminate the bath. This disadvantage was known to
Le Chatelier,* who sought to overcome it by enclosing the
anodes in porous cells. Although DeviUe's and Le Chatelier's
inventions were primarily intended to be implied to aluminium
reduction, they are equally available in the extraction of mag-
nesium, as Berthaut has recognised, for he has included the
reduction of both these metals in his patent, and has adopted
Deville's proposal as his own invention. Everyone, however,
who has experimented with electrodes of this description has
Fig. 3. — Gorup-Besanez apparatus.
become conWnced of their uselessness. Although the idea is
undoubtedly sound, it has not been possible to carry it into
effect owing to technical difficulties that are practically insur-
mountable.
Small-Bcale Apparatus. — A very simple apparatus for lecture
experiments has been described by v. Gorup-Besanez : t a clay
pipe is supported in a holder (Fig. 3), and the bowl is filled with
the double-chloride of magnesium and potassium, which is then
melted with the aid of a Bunsen burner. A knitting-needle is
now introduced through the stem until it comes into contact
* See AlumifUutn.
t Lehrbuch der anorQanischen Chem., 4th Ed., p. 517 (Gorup-Beaanex).
Digitized by VjOOQIC
6
ELECTRIC SMELTING AND REFINING.
with the fused salt. The needle being connected outside with
the negative pole of a battery forms the cathode ; a fragment of
coke connected with the other pole and dipped into the bowl of
the pipe from above serves as anode. After cooling, small
globules of magnesium are to be found distributed through the
solidified mass of salt; but the greater part of the magnesium
reduced in this experiment is burned.
Fischer's Apparatus. — Apparatus of the type suggested by
F. Fischer,* in 1882, in discussing other methods for the electro-
lytic decomposition of carnallite, have not come into use even
experimentally. An apparatus, also by far more suitable for
lecture purposes, was recommended by F. Fischer at a slightly
later date.t In this case the porcelain crucible (from the Royal
Porcelain Factory in Berlin) was enclosed within two cylinders,
a and 5, made of sheet iron lined with asbestos (Fig. 4), bound
Fig. 4. —Fischer's apparatus, 1884.
by three strong wires beneath, and supported on three feet, 2.
The object of llie rings was to equalise the temperature, so that
the crucible might be steadily and uniformly raised to a clear
red heat. The cover, g?, was also coated on the under side with
asbestos, and had an opening through which the crucible might
easily pass, in order that it should rest upon a thick iron wire,
protected from the action of the flame by a pipe-clay tube, x.
The hot gases from the three burners circulated evenly around
* Dingier* 8 polyttchnisches Journ.^ 1882, vol. ccxlvi., p. 28.
t Wagner-Fischer, JahrcfiUrichf der chem. Techv,, 1884, p. 1317.
Digitized by V^jOOQ IC
MAGNESIUM. 7
the crucible inasmuch as they were afterwards compelled [to
pass downwards between a and 6 in the direction indicated by
the arrows. As soon as the double salt was melted, a circular
asbestos plate, v, was placed upon the crucible, and pressed into
dose contact with the rim by a heavy cast-iron ring, /. Passing
through the asbestos cover was a clay tube, o (made by removing
the bottom from a porous battery cell), in the side of which a few
holes were bored. Through the clay tube there were passed
both the carbon which served as positive electrode, and a small
tube provided with a side connection, r, for the removal of
chlorine. These were kept in. place by means of asbestos plates.
This form of tube was chosen in order to facilitate the removal
of accidental obstructions, and to enable the operator, after
raising the stopper, to test for the evolution of chlorine by the
application of a strip of litmus paper. A length of iron wire, «,
5 mm. thick, served as the negative pole, and had its lower end
bent into a ring so as to encircle the carbon. A very gentle
current of reducing or neutral gas, previously dried by passing
through calcium chloride, was introduced through the tube ^,
and made its escape along with the chlorine through r. When
coal ga^ was used, the hydrogen and hydrocarbons present in it
combined with the chlorine.
The electric current was derived from a dynamo built by
Uppenbom and Gackenholz, and capable of affording a current
of about 50 amperes at from 9 to 10 volts; and the dynamo was
driven by a 1-H.P. gas engine made by Korting, of Hanover,
When one crucible was employed, the current-strength averaged
35 amperes ; but by coupling two crucibles in series, it ranged
from 22 to 25 amperes, which was equivalent to from 44 to 50
amperes in the two together, and the hourly out-turn of mag-
nesium amounted to 10 grammes [154 grains]. The metal was
deposited in the form of a slime upon the negative electrode, but
melted at a clear red heat into globules of the size of a nut
which slowly rose to the surface. Instead of the iron ring, e, a
carbon plate may be used as cathode, and from this the globules
of magnesium detach themselves more readily than they do from
the former. By this method the reduction of magnesium may
readily be made the subject of a lecture experiment.
For lecture and laboratory purposes the apparatus is certainly
useful ; but in no case is it adapted for regular use on a large
scale, as one would suppose it to be from the account given in
Wagner-Fischer's Ilandhuch der cJiemiachen Technologie (edition
1893), where it is the only form described. It could only be
made of comparatively small size, and even then would be very
fragile.
GraetzePs Apparatus. — A short time before the publication
of the above process, Graetzel's patent* was taken out. This
♦ German Patent 26,962, 18S3.
Digitized by VjOOQ IC
8
ELECTRIC SMELTING AND REFINING.
patent, which has been described very frequently, was a com-
bination of several ideas. The foUowing account is taken from
the specification : —
" The present improvements in the apparatus for the electro-
lytic production of the alkaline-earth metals on a manufacturing
scale from their respective chlorine and fluorine compounds,
Fig. 5.— Oraetzel's apparatus (transverse section).
F'ig. 6. — Graetzel's apparatus (longitudinal section).
with the subsidiary assistance of a current of reducing gas,
consist, on the one hand, in the use of the melting vessel itself
as the negative electrode, and, on the other hand, in the separa-
tion of the two electrodes in such a way that the chlorine
produced at the positive electrode can be carried off separately
Digitized by LjOOQ IC
MAGNESIUM. d
from the reducing gases. A dynamo is employed as the source
of electricity."
Of the accompanying illustrations, Figs. 5 and 6 show the
longitudinal and transverse sections respectively of the apparatus
as used for the production of magnesium ; Fig. 7 gives a longi-
tudinal section of the plant arranged for producing aluminium ;
whilst Fig. 8 is a view of the inner vessel, G.
The melting vessels, A, which also serve as electrolytic cells,
are set in the furnace, O, each in its own hearth ; they may be
arranged either in parallel or in series, and range from two to
five in number, according to the strength of current available.
The veasels. A, which may have any desired form, but are most
Fig. 7.
Graetzers apparatus.
Fig. 8.
conveniently crucible-shaped, are made of metal, and are used as
negative electrodes ; those intended for aluminium reduction are
of copper, iron, or steel, and those for magnesium of malleable
cast steel. Each stands upon a chamotte plate placed in the
middle of a grate, and the furnace is closed by means of a
chamotte cover, divided into two halves, and shaped to suit
the disposition of the melting vessel.
Each melting vessel is provided with a lid, e, e, of the same
metal as the vessel itself, and the reducing gases are led into
it from the main, o, through the tube, o^, and are aftenvards
conducted away through o^ into the return main, q.
In order to insulate the electrodes and to keep the chlorine
(which is evolved in K) separate from the reducing gas, the
carbon electi*ode is enclosed in a special vessel, G, which is
suspended within the crucible, A, through an opening in its
cover, e. This vessel, G, is made of chamotte, porcelain, or
other fire-resisting mateiial, which must at the same time
be a non-conductor of electricity, and is preferably cylindrical
Digitized by VjOOQIC
10 ELECTRIC SMELTING AND REFINING.
in shape. It is closed above with a cover, through which the
carbon electrode is passed, whilst at the bottom or on its sides
it is provided with perforations, g, which serve to give the
melted charge free access to the carbon anode. The chlorine
generated in 6 escapes through a side tube, p\ at the top of
the compartment in the chlorine main, pj which, like the
mains o and q, serves all the crucibles in the installation. The
coupling of several melting vessels into a battery is effected
in the usual manner, as indicated in the illustrations, where
m and n represent the connections with the dynamo.
For the reduction of aluminium the arrangement of the
apparatus is somewhat modified. This metal tends to sink to
the bottom of the melting vessel, and so to come into contact
with the negative electrode, which should, therefore, consist
of a separate metallic conductor (preferably of aluminium) in-
dependent of, and inserted within, the crucible. Accordingly,
the crucible, 8 (Fig. 7), is made of porcelain, stoneware, or other
equally refractory material, and is protected from direct contact
with the flame by an outer sheathing of metal. The metallic
conductor, r, placed within «, is connected to the negative pole
of the dynamo. If preferred, the cathode-cell may be introduced
into a graphite crucible instead of into a porcelain vessel, in
which case the outer metallic sheath may be omitted ; or it may
be used in conjunction with a metallic vessel, which must, of
course, be capable of resisting the heat to be applied.
In order to lessen the internal resistance of the apparatus,
and, at the same time, to enrich the bath as the supply of
electrolyte becomes exhausted, the plates or rods, M, which
consist of a mixture of alumina and carbon for the reduction
of aluminium, or of magnesia and carbon for that of magnesium,
are introduced into the compartment, G, parallel with the
carbon electrode, but quite independent of it. The carbon
combines with the oxygen of the oxide in these rods, and thus
enables the metal previously combined with the oxygen to
unite with the chlorine that is present in the compartment.
The following claims are made in this patent specification : —
1. An apparatus for the production on a manufacturing scale
of the earth metals from their anhydrous chlorine and fluorine
compounds, consisting of a closed metallic melting vessel, A,
which serves also as the negative electrode, with inlet and
outlet for reducing gas, in combination with the non-conducting
and fire-resisting vessel, G, which surrounds the positive electrode,
K, and is provided with openings in the sides or at the bottom,
to insulate the electrodes from one another, and to facilitate the
separate removal of the halogen produced. 2. In the manu-
facture of aluminium on a large scale in the above-described
manner, the use of metallic conductors, especially of aluminium
introduced into the melting vessel as negative electrodes, as
Digitized by V^OO^ ICl!
MAGNESIUM. 11
described with reference to Fig. 7. 3. To diminish the electric
resistance within the apparatus first described, as well as for
the enrichment of the smelting bath as it becomes impoverished,
the use of plates or rods, M (Figs. 6 and 7), of alumina and
carbon, or of magnesia and carbon, in equivalent quantities
respectively, which should be placed within the vessel, 6, in-
dependently of the electrode.
The only cuUual novelty in the whole patent was the itUro-
duction of the plates or rods of magnesia and carbon, or of
alumina and carbon, to be used in the production of magnesium
and aluminium respectively, and this hSw not proved successful ;
its use was discontinued in the aluminium and magnesium
factory at Bremen even at the time that Graetzel was managing
the works. The effect of these plates, like that of Deville's oxide-
carbon anodes, was only to introduce impurities into the bath,
and thus to cause irregularity and inconvenience in the working.
In order to bring about the desired reaction, as formulated in
the equation
MgO + C + Clj = MgCla + CO
the temperature of the bath would have to be maintained much
higher than is desirable either for the production of magnesium,
for the preservation of the apparatus, or for preventing the
destruction of the plates themselves through the removal of
the oxide.
The introduction of reducing gases into the cathode chamber
was made a prominent claim in the specification. But this pro-
posal had been made in the year 1882 by F. Fischer.* Hiller,t
also, had arranged for the use of a reducing atmosphere in the
cathode chamber of his apparatus for the reduction of strontium
and lithium. Moreover, this precautionary measure is entirely
unnecessary in the electrolysis of camallite. Practically the
whole of the magnesium remains adhering to the immersed
electrode and to the walls of the crucible, provided that a
current of suitable strength be used and that the fused salt be
not connected up with the battery until it has been melted
sufficiently long, and at a sufficiently high temperature, to
drive off the last traces of water, which are retained by it with
the most obstinate pertinacity. The use of reducing gas is not,
therefore, necessary on this account; but if the crucible be
immersed in the fire almost to the top rim it will be found that
in the absence of reducing gases the walls of the vessel become
strongly corroded above the level of the fused mass within.
This is due to the metal of the crucible becoming chloridised,
owing to the combined effects of the hot fire-gases without and
the acid gases within, the latter being always present above
the surface of camallite when it is melted in the presence of
•See footnote, p. 6. t F. Hiller, Lehrhvch der Chemie, 1863.
Digitized by VjOOQ IC
12
ELECTRIC SMELTING AND REFINING.
•even a small proportion of air. The crucible walls, therefore,
above the level of the fused substance become rapidly cor-
roded through, whilst, in addition, the melted camallite tends
to creep over the rim of the vessel on to the outer surface,
and there, aided by the furnace gases, it exerts a most de-
structive influence. This action may, however, be prevented
by adopting the construction of crucible used by Borchers,
This crucible (Fig. 9) is provided, at a distance of about 2 to
2^ ins. from the top, with a flange, which serves to support it
with its lower portion immersed in the fire. The upper
part of the crucible being thus exposed freely to the air
remains sufficiently cool to reduce to a minimum both the forma-
tion of acid gases and the corrosion
of the vessel. Thus the furnace gases
come in contact only with Uiose
portions of the crucible walls which
are in contact with fused substances
within, and which serve as cathode
surfaces. The tendency of the melted
salts to creep up the side of the
crucible is completely prevented be-
cause, even if the uppermost margin
of the fused mass within the crucible
should not be crystallised in contact
with the comparatively cool walls,
yet any thin layer of fused salt that
might rejich above this level would be at once solidified.
Finally, the use of the melting vessel as cathode, claimed by
Oraetzel as his invention, was anticipated in actual practice by
Davy so long ago as the year 1808.
Napier's Electrolytic Vessel. — An account may be given
here of an apparatus patented by Napier* in 1844. It was,
indeed, originally intended for the extraction of copper, but
we find combined in it, even at this date, the most important
elements reappearing in GraetzeFs and other patents afterwards.
Xapier used a large crucible or other convenient vessel made
of some conducting material, of which the inner side, with the
exception of the bottom, was lined with a coating of clay. The
copper ore, roasted as free as possible from sulphur, was mixed
with the necessarv fluxes and fused in the crucible ; the melted
ma«s was then submitted to the action of the galvanic current in
such a way that the crucible itself formed the cathode, whilst
the anode was an iron rod united to a plate at the lower end.
It is thus clear that even in 1844 an apparatus for the reduction
of metals had been patented,! which consisted of a melting vessel
Fig. 9. — Borchers' flanged
crucible for electrolytic
reduction.
♦ English Patents 10,362, 1844, and 684, 1845.
Frankhn JiiAt., 1889, vol. cxxv., p. 376.
+ C/. Graetzel's Patent claims, pp. 10, 11.
Cf. Houston, Jottm,
Digitized by VjOOQ IC
MAGNESIUM.
IS
made of some conducting inaterial and serving as the negative
electrode, in combination with a non-conducting and fire-resisting
vessel, open at the bottom, and surrounding the positive electrode
for the purpose of insulating the electrodes and facilitating the
removal of the anion.
Borchers' Magnesium Process. — In accordance with the
principles explained above, and following the pattern of Troost's*^
al^
Fig. 10. — Borchers' experimental furnace for the electrolysis
of fused salts.
apparatus, the cell surrounding the carbon electrode being re-
duced in size, whilst at the same time taking into account the
experience which led to the adoption of the shape of crucible
depicted in Fig. 9, the author has designed an apparatus (Fig. 10)
suitable for a large number of experiments in the electrolysis
of fused salts. Within the iron crucible, K, which forms the
cathode, is suspended the carbon rod, A, as anode, surrounded
• See p. 22.
Digitized by VjOOQIC
u
ELECTRIC SMELTING AND REFINING.
by the porcelain tube, C. The anode is connected to the con-
ducting wire from the dynamo by means of the screw clamp, V,
and is supported by the ring-shaped porcelain cover, L, whilst
the porcelain tube rests by means of a collar on another annular
porcelain plate, d ; the weight of the whole crucible is in turn
borne by the chamotte-plate cover, D, of a Perrot furnace, upon
which it is held by the flange, F; and this, by an extension on
one side, makes electrical connection with the negative pole of
Fig. 11.— Borchers' furnace for use with currents of 100 amperes.
the dynamo through the binding screw, N. The furnace consists
of a wide chamotte cylinder, O, protected externally by a sheet-
iron jacket, M, which may be supported either by feet attached
to it, or by a temporary stand ; and the whole is closed beneath
l)y a chamotte plate, B, provided with a central opening. The
internal fireclay cylinder, W, serves to convey the heated gases
from a sufficiently powerful gas burner, first of all upwards and
around the crucible, and thence downwards through the annular
Digitized by VjOO^ li:!
MAGNESIUM.
15
space between W and O to the flue, Z. The upwardly bent end
of the flange, F, with the negative main, N, is connected by
means of a screwed joint.
In using the apparatus, the empty crucible and iU^ appur-
tenances should first be heated for a considerable time, while
the camallite is being melted, preferably, in a second crucible.
Whilst the crucible just described is suitable for currents of
about 50 amperes, it is possible to employ 100 amperes with the
apparatus used in Borchers* laboratories in the Technical College
at Aachen, Fig. 11. This apparatus is heated by means of a
Fig. 12.— Borchers* furnace for the electrolysis of magnesium
compounds.
Rossler gas furnace. An arrangement adapted to coal firing
and to a current of 250 to 300 amperes is shown in Fig. 1 2.
It is not advisable to use an apparatus much larger than that
which is here figured, not only for structural reasons, but on
account both of tiie size of the electrodes that would be required
and of the distance between them. It is preferable to use a
greater number of vessels of ordinary size connected up in series ;
but in that case the crucibles must be supported in separate
furnaces, in order to avoid short-circuiting between them through
the deposition of soot or other conducting material. Several
Digitized by VjOOQIC
16 ELECTRIC SMELTING AND RBFININO.
melting vessels, however, may be served with fuel from a single
generator. The grate, whether gas firing or solid fuel be used,
is best built in a separate chamber, so that the crucible may be
protected from the direct action of the flame (little extra heat
being required during the actual progress of electrolysis), and
that the fluid contents of a leaky crucible may be prevented
from flowing into the fuel. To facilitate the removal of such
material a small collecting chamber may be provided beneath
the crucible, as shown in the illustration.
It is necessary to arrange for a heating chamber, in which
spare porcelain or clay portions of the apparatus may be kept
hot, so that damaged parts may be replaced without loss of time.
This chamber is conveniently heated by the waste gases from the
furnace.
Preliminary Preparation of the Charge. — ^The first stage
in the extraction of magnesium on the manufacturing scale is
the production of anhydrous camallite. The salt as it occurs
in nature is too impure for immediate treatment by electrolysis ;
hence the camallite necessary for the production of the mag-
nesium is prepared from magnesium chloride and potassium
chloride, the latter of which is in part recovered after the
exhaustion of the magnesium in the fused bath under electro-
lytic treatment. Under a new patent of the Hemelinger
Aluminium and Magnesium Works,* natural camallite can be
rendered serviceable for electrolysis by fusing it, without any
previous purification, in admixture with so much chloride of
magnesium or chloride of sodium, or of dehydrated artificial
camallite, as may be necessary to bring it up to the percentage
composition, MgCl, = 41-66, KCl = 32-66, and NaCl = 25-66,
corresponding to the formula MgClo . KCl . NaCl. A fused
bath of this composition is well suited for electrolysis, and
if a little fluorspar be added gives a metal which flows readily
together.
By whichever method the mixture of salts is prepared, the
following procedure is usually adopted: — Crystallised magnesium
chloride, or artificial camallite, is fused in an open hemispherical
iron pan, either of these salts melting with comparative ease in
its water of crystallisation. The necessary quantity of potassium
chloride, or of the mixture specified in the Hemelinger patent,
is then added to the fluid mass, which is at first pasty, but
which rapidly becomes solid again as it loses its water of crystal-
lisation. During the intermediate pasty stage it is necessary to
stir the mass briskly with an iron tool, both to facilitate the
drying of the charge and to prevent the adhesion of thick crusts
of salt to the hotter portions of the walls of the pan. It is well
known that without the addition of the salts above mentioned,
hydrated magnesium chloride would, when evaporated in the
♦ German Patent 116,015.
Digitized by LjOOQ IC
MAGNESIUM. 17
ni&nner just described, be converted for the most part into
magnesium oxide, or oxychloride, and hydroohloric acid, as
shown by the following equation: —
M^Cla + HgO = MgO + 2HC1.
But this reaction is not entirely prevented even by the presence
of the alkali-metal chlorides, so that fumes of hydrochloric acid
escape during the operation, and these must be conducted away
by means of a hood.
Even after the temperature has been raised to from 300"* to
400* C, at the end of the operation, the mixture is not absolutely
anhydrous, but it is ready for fusion, whereby the last traces of
water will be expelled. This fusion is effected in vessels similar
to those used for the electrolytic process, these vessels are sus-
pended in the same way above a firing chamber, from which they
can be readily removed. A temperature of from 500** to 600** C.
is necessary for fusion. During the process a portion of the
magnesium chloride reacts, according to the equation above
quoted, with the residual water vapour, which is only finally
expelled at a red heat, so that the result of the dehydrating
process is the production of a fused mass containing a more or
less considerable proportion of magnesium oxide. This oxide
is, during electrolysis, a great hindrance to the running together
of the globules of magnesium formed during the process; and
the greater the quantity of oxide present in the fused mass, the
greater will be the proportion of spongy metal which can with
difficulty be recovered in useful form. It is therefore necessary,
before starting the electrolysis, to reconvert into chloride any
magnesia which has been formed, either in the manner above
described, or by another reaction which will be referred to in
the next paragraph. This may be most readily done by the
process already described by Bunsen, in which ammonium
chloride is added to the mass in the form of sublimed sal-
ammoniac, thus : —
MgO + 2NH4CI = MgCLj + 2NH, + HjO.
Oettel* has shown by experiment that the presence of mag-
nesium sulphate in carnallite is disadvantageous, inasmuch as
the magnesium reduced electrolytically tends to react in part
with the sulphate, with the production of magnesia and sulphur
dioxide, thus : —
MgS04 + Mg = 2MgO + SO,.
Owing to the fact that this reaction takes place very gradually,
a larse number of the minute globules of magnesium become
coated, as fast as they are reduced, with a thin film of mag-
♦ Zeitachrift/ur EUktrochemie, 1895, vol. ii., p. 394.
2
Digitized by VjOOQ IC
18 ELECTRIC SMELTING AND . REFINING.
nesium oxide, which checks their uniting together into larger
masses. Oettel, as a simple remedy for this difficulty, recom-
mends the addition of a small proportion of carbon to the
carnallite at the time of fusion. This carbon, which may be
in the form of wood-charcoal dust, sawdust, sugar, flour, or
the like, decomposes the magnesium sulphate at a dull red
heat, as follows: —
MgS04 + C = MgO + CO + SO,.
The Beduotion ProoeBs. — ^After the removal of the sulphate
originally present, and of the oxide formed during the fusion of
the mixture of salts, the fused mass will, at a clear red heat,
soon become tranqidl and clear. When this point has been
reached, the melting pan is lifted out of its seat and its contents
are emptied into a pre-heated electrolytic vessel, taking care
that any sediment, consisting of impurities, is left in the melting
pan. The anodes, with their porcelain envelopes, which must
also have been heated beforehand, are then introduced, and the
process of electrolysis is started.
If all the precautions have been properly taken, the magnesium
will be found to deposit on the lower part of the walls of the
electrolytic vessel in globules, which increase rapidly in size,
while the chlorine, rising through the porcelain cylinder sur-
rounding the anodes, escapes above. It is most important that
the temperature of the bath be not permitted to rise much above
that of the melting point of magnesium, because at this tempera-
ture the metal is specifically heavier than the electrolyte, and
will therefore remain at the bottom of the vessel. A higher
temperature of bath is favourable to the re-solution of mag-
nesium, and the pressure required will be sensibly higher; it
leads also to a loss of magnesium, because the metal at a bright
red heat is specifically lighter than the fused salt, and so tends
to float up from the bottom, in which case a considerable propor-
tion of the magnesium will be burned, either at the surface or at
the anode.
Working with a current-density of at least 1,000 amperes per
square metre [93 amperes per sq. foot] of cathode surface, the
current-density at the surface of the anode will be about ten
times as much, even if an unusually thick carbon rod be employed.
In spite of this, a pressure of only about 7 to 8 volts is required,
whidi may be reduced by 1 or 2 volts (where an economical
installation is sought) by increasing the size of the anode.
After continuing the electrolysis for a sufficiently long time,
an inspection of i£e contents of the crucible, through the dear
fused salt, will show when the required quantity of the metal
has been accumulated. The current is then broken, the screw
connections are opened, and the cover, d, together with all that
it supports, is removed from the melting vessel. The flame is
Digitized by LjOOy ICl!
MAONiGSlUM. 19
now somewhat increased, and the mass of metal clinging to the
walls is detached with the aid of an iron scraper shaped to
correspond with the inner surface of the crucible. The whole
contents of the vessel are then poured into a flat box made of
sheet iron, which must be cold and perfectly dry, and any metal
that still adheres to the walls of the vessel must be rapidly
scraped off. The solidified mass, after cooling, is broken up,
and the globules of magnesium are picked out. The larger pure
shot may be melted together without flux in a plumbago crucible,
but the less pure metal must be fused and refined.
Melting of Eleotro- deposited Magnesium. — ^The larger
and purer pieces of magnesium may be melted together in a
crucible in an ordinary fire, but the less pure and the fine-
grained material must always be submitted to a refining opera-
tion. For this purpose carnallite is melted in an iron crucible,
and the crude magnesium is thrown into the fused mass. At a
dull red heat an iron rod is introduced, and with its aid the
metal lying at the bottom of the crucible is as far as possible
pressed into one mass. The temperature is now gradually
raised to that of a bright red heat, until a point is reached at
which the specific gravity of the magnesium is less than that of
the flux. The metal then, liquating from its impurities, floats
to the surface in the form of globules of considerable size, and is
removed in perforated sieve-like ladles. The sur&M^tension of
melted magnesium is so great that the metal cannot flow through
the fine perforations of the ladle, although the flux may thus
be drained completely away. The purified metal is finally
melted together in an iron crucible to separate the last traces
of slag, and to enable it to be cast into the bars or rods required
for the market.
Electro-thermal Reduction of Magnesium. — Knofler and
Ledderboge**^ have proposed to smelt magnesium by placing a
mixture of carbon and magnesia in the form of rods in an electric
circuit after the manner of carbons in an arc lamp ; the
magnesium and carbonic oxide thus produced were to be pre-
vented from reacting with one another, either by the introduc-
tion of reducing gas or by conducting the whole experiment
in vacuo. The process would probably be workable, but it is
questionable whether it would be successful financially. The
considerable electromotive force required for the operation, the
high price of the magnesium-unit in pure magnesia as compared
with carnallite, the circumstance that magnesium is volatile at
temperatures much below that required for the decomposition
of ihe oxide, and the difficulty of condensing the resulting
metallic vapour from its admixture with carbon monoxide and
other gases, are sufficient grounds for predicting the practical
&ilure of the process.
♦ Gorman Patent 49,329, Feb. 6, 1889.
Digitized by VjOOQ IC
20 ELECTRIC SMELTING AND REPINING.
Hilberg's Process. — The means recommended by Hilberg*
for improving both the process itself and the yield of metal
possess no advantages. He proposes to cover the sur&u^ of the
fused mass with a layer of asbestos cloth, with the object both
of protecting the electrodes from blows or shocks of any kind,
and of lessening the frothing of the bath. The asbestos is also
intended to prevent the atmospheric oxidation of the electrolytic
magnesium, and to promote this object still further the space
above the fused mass is to be rendered vacuous, or to be filled
with an indifferent gas. From what has been written above
concerning the process, it will be obvious to the practical man
that it is not necessary to introduce the complications of the
Hilberg patent in order to overcome the difficulties of the
process.
Applications of Magnesium. — I'he use of magnesium haa
been considerably extended of late, since the discovery by Mach
of the valuable properties of its alloy with aluminium. Besides
the Hemelinger Aluminium and Magnesium Works, near
Bremen, the Griesheim-Elektron Chemical Works is also pro-
ducing magnesium, by a later process worked out by Bathenau
and Suter, at the electro-chemical works at Bitterfeld, and kept
most rigorously secret.
Otherwise the use of magnesium remains restricted, and the
hopes, at first entertained, that it might be utilised in the manu-
facture of aluminium have remained unfulfilled, owing to the
introduction of more advantageous methods for the latter pur-
pose ; its employment, however, as the source of a powerful
light in pyrotechny and photography is by no means incon-
siderable. The price of magnesium, as compared with that of
aluminium (at the present time) and the alkali metals, especially
sodium, has led to its introduction into the chemical industries
only in very limited quantities, in spite of its high reducing
power. It is, however, especially well suited to the dehydration
of alcohols, ethers, and oils, because the hydroxide formed by
its reaction with water is quite insoluble in most of these
substances. If it be still employed, as proposed by Fleit-
mann, in melting nickel to remove the last traces of oxide
dissolved in the metal, it is less on account of its reducing
properties, which it possesses in common with aluminium, than
because any excess of magnesium that may be added does not
alloy with the nickel. Magnesium is also recommended for
the refining of other metals, such as copper, German silver, and
steel, being added to effect the removal of small quantities of
dissolved oxides, sulphides, and phosphides. In laboratories it
is employed as a powerful and reliable reducing agent.
* English Patent 16,659, of July 30, 1898, and German Patent 110,403,
of Feb. 5, 1898.
Digitized by VjOOQ IC
LITHIUM. 21
CHAPTER 11.
LITHIUM.
Ocourrenoe of lathium in Nature. — Of the compounds of
lithium which occur in nature, the chloride is found in many
mineral waters ; a fluoride together with a silicate is met with
in lepidolite and a phosphate in triphylite. Numerous other
minerals, together with a few plants {e,g,^ tobacco), also contain
lithium, but even the richest minerals never contain more than
4 per cent, of this metal, and usually only from 1 *5 to 2 per cent.
Properties of the Metal. — Lithium (Li ; atomic weight = 7 ;
specific gravity = 0*5936) is a soft metal belonging to the alkali*
metal group. Surfaces of the metal protected from the air have
a white silver-like lustre. It melts at 180® C, and has been
vaporised by Troost at a clear red heat (about 1,000* C). At
temperatures below 200' C, it may be melted in the air without
taking fire, but at higher temperatures it bursts into flame,
emitting a blinding white light, and becoming converted into
oxide with the evolution of much heat. It unites with sulphur
and the halogens under similar circumstances, evolving both
light and heat. It decomposes water at the ordinary tempera-
ture, and floats to the surface where, unlike the other alkali
metals, it remains tranquil, without melting, giving rise to flame,
or lighting the hydrogen produced by the reaction. It is obvious
that the action of acids must be very energetic upon a metal
such as lithium, which is able to decompose water, and the salts
of which are, almost without exception, so easily soluble.
Extraction of Ijithimn. — ^The many methods proposed for
the treatment of these minerals are so far alike that the lithium
is in all cases rendered soluble by converting it into the sulphate
or chloride. As a natural result of this treatment, it follows that
in addition to the more easily separable metals, the alkali- and
alkaline-earth metals, which are always present, must pass into
solution witih the lithium. The separation of these latter salts
at present introduces great difficulties, and adds considerably
to the cost of the process. Such an operation is, however, un-
necessary if it be intended to produce metallic lithium, because
this metal may be completely separated in a perfectly pure con-
dition and with the greatest ease (in the absence of magnesium
salts) by electrolysing a fused mixture of alkali-metal and al-
kaline-earth metal chlorides. Bunsen and Matthiessen in 1 854,
Digitized by V^jOOQ IC
BLBCTBIC SMELTING AND REFINING.
working with lithium chloride, showed that the successful elec-
trolysis of fused alkaline chloride was quite possible ; and the
former* wrote as follows on this point: — "The lithium chloride
is melted in a small thick-walled porcelain crucible by means of
a Berzelius lamp, and is then treated with the current from 4 to
6 zinc-carbon elements. The current is passed through the fused
chloride from a pointed fragment of gas carbon to a piece of iron
wire the thickness of a knitting needle. After the lapse of a
few seconds only a molten silvery- white regulus is seen to be
forming upon and clinging to the im-
mersed portion of the wire ; and within
two and a-half to three minutes it will
have grown to the size of a small pea.
The metal is obtained, with the aid of a
small spoon, by raising the melted
regulus out of the solution, together with
the wire electrode ; and as this operation
may be repeated at intervals of three
minutes, it is an easy matter to reduce
an ounce of lithium chloride in quite a
short time."
Troost's Prooess. — The publication
of Bunsen's results led naturally to
further work in the same direction. In
publishing his confirmation of these re-
sults, Troost,t in 1856, described a modi-
fied apparatus which, in view of the
patent claims advanced by later inven-
tors, is worthy of special attention. A
cast-iron crucible, T (Fig. 13), 4f inches
high and 2 inches wide at the mouth, was
used as the melting vessel. This was
covered by a close-fitting lid, D, provided with two perforations,
one of which was 0*2 inch wide, and served to admit the negative
pole, K; whilst through the other, which was IJ inches in
diameter, there was passed a sheet-metal cylinder, having an
internal diameter of 1-14 inches, and reaching downwards to
half the depth of the crucible. Within the metal cylinder was
a porcelain tube forming a sheath for the positive pole. A, and
a means of escape for the separated chlorine. The lithium
gradually accumulated at the negative pole, and the apparatus
could be left to itself for an hour at a time, provided only that
the decomposed chloride was replaced from time to time, as
required, by the introduction of fresh salt through the porcelain
tube.
• Lithig'sAnn.j 1855, vol. xoiv., p. 107.
t Comptes Rtndusy 1856, vol. xliii., p. 921 j and ^im. dt Chim. et de Phys.^
1856, vol. li., p. 112.
Fig. 13.
Troo8t*8 apparatus for
the electrolysis of fused
lithium chloride.
Digitized by LjOOQ IC
LITHIUM. " 23
According to the author's experience with apparatus of the
kind just described, lithium and other metallic chlorides may be
electrolysed in them for a short time. Even, however, if the
negative electrode be initially insulated from the iron cover and
crucible, the sepai-ated metal floats to the surface and makes
electrical connection with the crucible and the sheet-iron tube,
so that all the iron portions of the apparatus which are below
the level of the fused salt soon form part of the negative pole.
So far as the separation of lithium is concerned, this action is in
itself harmless ; but, unfortunately, the metal becomes deposited
on the inner side of the sheet-iron tube, and as lithium very
readily attacks both silica and alumina, the porcelain tube with
which the metal is now in contact is so attacked that it is soon
rendered porous, and the inner surface of the iron then becomes
the principal seat of reduction for the lithium. The result of
this is that the electitxles are insufficiently separated, and short-
circuiting is brought about through the porous substance of the
porcelain tube. Such an apparatus may not, therefore, be used
continuously for any length of time. There can be no doubt
that in the arrangement, as Troost described it, the crucible
either acted as the negative pole from the very beginning of the
electrolysis (for Troost does not mention that it was insulated
from the cathode wire), or, if it were insulated at first, that it
did so within a very short time afterwards, for the reduced
metal rapidly accumulates upon the surface of the liquid in
sufficient quantity to bridge over the space between it and the
original wire electrode.
Accordingly, in 1866, we find Troost* using "for the produc-
tion of metals from their anhydrous chlorine compounds, a closed
metallic vessel, which serves also as negative electrode, in com-
bination with a non-conducting and fire-resisting vessel, which
is open at the bottom, and surrounds the positive electrode, with
the object of separating the two electrodes and of conducting
away the chlorine produced."
Hiller's Apparatus. — Hillerf has described an apparatus for
the production of strontium that is especially well adapted to
extraction of lithium from its chloride.
In Fig 14, t is a porcelain crucible which contains the fused
lithium chloride. The cathode consists of an iron wire, of which
the end. A;, is enclosed in the bowl of a clay tobacco pipe, p. The
lithium separates at k, and, being specifically lighter than the
fused chloride, collects upon the surface of the liquid in the
pipe-bowl, which protects it completely from the re-oxidation
that would otherwise result from contact with the air. It is
advisable, before commencing the electrolysis, to remove the
air from the pipe. To effect this, a glass tube, g, is attached
* Compare GraetzeUs claimB in his patent Bpecifioation, on pp. 10, 11.
t F. Hiller, Lehrbuch der Chemie, Leipzig, 1863.
Digitized by
Google
24
ELECTRIC SMELTING AND REFINING.
to the pipe-stem by means of a perforated cork ; beyond this is
a narrow glass tube, rf, leading to a caoutchouc tube, e. The
second glass tube is fitted to the first with the aid of a cork at c,
and between this cork and the outer tube is passed the iron
electrode wire. The wire is wound several times around the
small glass tube at d, and when the end, k, has been adjusted in
the centre of the pipe-bowl, the whole arrangement is held
firmly in position by covering the joint, c, with plaster of Paris.
The air in the pipe is displaced by passing a current of per-
fectly dry hydrogen through the caoutchouc tube, e ; the bowl
of the pipe is then dipped into the fused chloride, the caoutchouc
Fig. 14. — Uiller's apparatus for the extraction of liihinni.
tube is disconnected from the hydrogen generator, the bowl is
sunk into the chloride to the depth indicated in the figure, and
the rubber tube is finally closed with a pinch-cock, q. It would,
however, be still simpler to substitute hydrocarbon vapour for
hydrogen ; to this end it would only be necessary to introduce
some paraffin into the upper part of the pipe-bowl. On immers-
ing the bowl in the melted salt, the paraffin would vaporise and
drive out the air before it through the caoutchouc tube, f, which
could then be closed by the clamp as before. In order to prevent
the contact of the separated lithium with the walls of the pipe,
which would lead to the contamination of the metal with silicon,
the inner side of the bowl is coated with a thin layer of graphite.
The graphite is first made into a thick paste by stirring it well
with a dilute solution of lithium chloride; the paste is then
painted over the surface of the pipe-bowl, and after it has become
Digitized by VorOO^ It!
LITHIUM.
25
thoroughly air-dried, the whole ia heated to a red heat. The
anode, a, consists of a gas-carbon rod connected to an iron wire,
€k. From 3 to 4 Bunsen cells are sufficient for the electrolysis.
As soon as the circuit is completed decomposition sets in, which
may be recognised by the copious evolution of chlorine at the
anode, and it is now only necessary to see that the lithium chloride
remains in a state of fusion. At the end of an hour the current
may be interrupted, and the fire drawn from the grate ; then,
after cooling completely, the crucible and pipe may be broken,
and a regulus of lithium will be found surrounding the iron wire.
Although this apparatus might easily be made on a larger
scale, and although it represents a distinct advance towards the
economical electrolysis of alkali-metal chlorides, yet it does not
satisfy all the requirements of a profitable installation.
Grabau's Apparatus. — The electrode cell patented by
Grabau,* for use in the extraction of metals of low specific
Fig. 15. Fig. 16.
Grabau's electrode cell for use with light metals.
gravity, may here be described on account of its similarity to the
last-named apparatus. The bell-shaped cell, a (Figs. 15 and 16),
is suspended by the stem, 6, from the transverse support, rf, and is
provided with the pole,^ which terminates at the highest point
of the inner surface of tiie bell. The cell, a, is closed above, but
it is open below, and should be quite immersed in the bath.
Either by a siphon-shaped tube, or otherwise, the air initially
present is given a passage for escape so that the bell may be
completely filled with the fused material. The metal separates
on the pole, f. But this apparatus cannot be used for an extended
operation, for there is no non-conducting material known that
will, under these circumstances, afford resistance for any length
of time to the action of the alkali metals in the presence of
fused alkali-metal chlorides. Even a few hours will suffice to
bring about the perforation of strong and thick porcelain tubes
• Cicniian Patent 41,494. 1887.
Digitized by VjOOQ IC
26 ELECTRIC SMBLTINO AKD REFINING.
by the alkali metals, in consequence of the reduction of the
silicates by them.
Guntz's Prooess. — A notable ignorance of past work in prac-
tical electro-chemistry is shown by Guntz,* who, in December,
1893, presented to the French Academy the description of a
process, from which the following account is taken: — A mixture
of 200 to 300 grammes of potassium and lithium chloride in equal
proportions (fusing point 450" C.) is melted over an ordinary
Bunsen burner. When fusion is complete the electrodes are
introduced. The anode consists of a carbon rod 8 ram. in dia-
meter, the cathode of an iron wire about 3 to 4 mm. thick. The
latter is enclosed within a glass tube 20 mm. wide. After
applying a current of 20 volts and 1 0 amperes for abotU an hour
there will be found in the glass tube a fluid column of lithium,
which stands at a level of more than 1 cm. above that of the
fused salt. To obtain the metal the electrodes are separated, the
glass tube is raised out of the liquid, and the fluid lithium which
floats upon the surface of the melt is removed with the aid of an
iron spoon and poured into a dry ingot mould.
All this, however, had been accomplished thirty years pre-
viously, only the earlier experimenters had been more economical
of electric energy, of which they had used but one-fourth part
of that employed by Guntz.
As has already been shown, the conditions required for the
extraction of liUiium are of the simplest, as compared with
those to be observed in the case of the better-known alkali
metals.
Treatment of Lithium Minerals. — In the treatment of
lithium minerals, a solution is first obtained which contains only
the chlorides of the alkali- and alkaline-earth metals. This solu-
tion is then evaporated to dryness in iron vessels, and it should
be kept slightly alkaline in order that it may not take up much
iron from the pans in which it is being boiled. A small quantity
of iron salt in the resulting chloride is not objectionable, how-
ever, for it is decomposed at the very outset of the subsequent
electrolysis, and the reduced iron does not alloy with lithium.
A little sal-ammoniac is now added to neutrali.se any free alkali,
and the mass is fused for electrolysis in an iron vessel, which
may, if desired, be the one that shall serve as cathode. The
arrangement of the electrolytic plant recommended for use in
magnesium or sodium manufacture may be employed. If the
former (p. 12) be selected, the rim of the crucible which projects
above the flange should be surrounded with a metal tube through
which water is circulated, in order to keep it cool and so to
produce a thin layer of crystallised salt upon the sides of the
crucible at the upper level of the chloride bath. Globules of
lithium, which might otherwise creep up the side, are retained
• Comptes Hendusy 1893, vol. cxvii., p. 732.
Digitized by V^jOOQ IC
LITHIUM. 27
by this layer of solidified salt. The electrolysis must not be
oontiniied so long that the lithium forms a metallic connection
between the walls of the vessel and the porcelain sheath of the
anode carbon, otherwise the same difficulties will occur that
were explained in the description of Troost's apparatus. After
a sufficient quantity of lithium has accumulated, the method of
procedure may be exactly the same as in the extraction of
magnesium already described; but there will necessarily be a
greater loss of metal in pouring out the contents of the vessel in
tiie case of lithium than in that of magnesium. In order to
prevent this loss, the whole charge may be cooled in the melting
vessel, but then it is important that the latter should be made
conical in shape. The crystallised salt loosens itself in solidify-
ing, and may be easily separated from the walls of the crucible
by blows from a hammer applied to the outer surface. The
shots of metal adhere firmly to the inner wall of the vessel.
They are removed by means of knives or scrapers, and separated
from adhering salt by melting them together in a paraffin bath
which is kept at a temperature of from 180° to 200° C; the
metal then floats to the surface of the bath, whilst the salts
sink to the bottom. From this bath the lithium is removed in
perforated ladles ; it is afterwards washed in benzene, and then,
after re-melting by itself, it is either cast in the shape of small
rods, or it is pressed into this form under gasoline (sp. gr,
0*66). It is finally stored in glass tubes hermetically sealed.
Kahlenberg's Process. — A method of great scientific interest
for the reduction of lithium has been discovered by Kahlenberg,*^
but it can scarcely prove to be of practical use, owing to the
unsuitable properties of the solvent employed for the electrolyte.
Lascynski and Gorski f had already, in their determinations
of conductivity, found that lithium chloride was soluble with
electrolytic dissociation in pyridine ; and Kahlenberg succeeded
in depositing this metal from a concentrated solution of these
compounds, using a carbon plate as anode, and employing a
current-density of from 20 to 30 amperes per square cm. .
[0-013 to 0*19 ampere per square inch] at a pressure of 14
volts, the lithium under these conditions being obtained in a
dense, adhesive, silver-white, metallic form.
In a permanent installation the chloride would be electrolysed
with the aid of the apparatus described in the chapter relating
to sodium, as suitable for the reduction of that metal ; but in
this case it must be remembered that the current-density should
never be so high as is employed for the extraction of sodium*
The deposition of lithium is efifected by a current of about
1,000 amperes per squai-e metre [0*64 ampere per square inch]
of cathode surface, at a pressure of about b volts.
* Jotim. of Phys. Chem,, vol. iii., p. 602.
t Zeitschr./Ur Elektrochem., 1897, vol. iv., p. 290.
Digitized by LjOOQ IC
28 ELECTRIC BMELTIKO AND REFINING.
The electrolytic method described, as well as the subsequent
purification of the metallic product, oflPers so little difficulty,
that the author considers this way of separating lithium from
the alkali metals for the purpose of obtaining pure lithium
compounds, to be simpler than the troublesome separation of
the salts in aqueous solution. The lithium obtained is abso-
lutely free from alkali- and alkaline-earth metals.
TTses of Ijithium. — Lithium has not found application in
the arts on account of its high price. The simplicity of the
electrolytic extraction process marks it out for use in the pro-
duction of pure lithium salts. The proposal to use lithium for
the generation of hydrogen in aeronautics (1 lb. of lithium affords
over 26-5 cubic feet of hydrogen at 0° C. and at the normal
atmospheric pressure) must still be relegated to the future.
Digitized by VjOOQ IC
SODIUM. 29
SECOND QROUP.
SODIUM AND POTASSIUM.
CHAPTER III.
SODIUM.
Occurrence in lE^'atnre. — Like all the alkali -metals and
alkaline-earth metals, sodium occurs in nature only in the form
of salts ; as chloride, in rock-salt, NaCl ; as fluoride, in cryolite,
AljFg . 6NaF ; as sulphate, in Glanber's salt, Na^SO^ . IOH.,0 ;
as nitrate, in Chili saltpetre, NaNOg; as borate, in borax,
Na^^O^ . lOHgO, and other borates ; as carbonate, in soda,
NajCOg.HgO, and in trona, (NaHCOgk . NagCOg . 2H2O ; as
silicate, in felspar, &c. For the manufacture of soda com-
pounds on the larger scale the chloride is mainly used; and
the metal may be obtained either by a metallurgical reduction
process from the carbonate or hydroxide, or by electrolysis from
the fused chloride.
Properties of the Metal. — Sodium (Na ; atomic weight =
23 ; specific gravity = 0*974), which is contained in these com
pounds, and which was first obtained by Davy from the
hydroxide, is a white metal, the freshly-cut surfaces of which
exhibit a silvery white lustre ; it is soft and may even be
kneaded at ordinary temperatures. It melts at 95*6** C, and
begins to vaporise at a clear red heat — i.e., at about 900* C. It
forms alloys with the other alkali metals, and with some of
the heavy metals. Of these alloys the amalgam, and the lead
and tin alloys, are of more or less importance in electro-chemical
practice. Anhydrous liquid ammonia is also a simple solvent
for sodium, the metal dissolving in this medium with a blue
colour. Molten haloid salts of sodium, especially the chloride,
have also the property of dissolving the metal, and it is stated
that a sub-chloride of the composition Na^Cl* is thus formed.
The fact of the solubility of sodium in its chloride is beyond
dispute, but there is as yet no decisive proof of the formation
of this sub-chloride; indeed, the phenomena attendant on the
electrolysis of fused sodium chloride point much rather to the
metal being in a state of solution, similar to that of gold in
• Boee, Poggendorff^s Anrialen, vol. xxxi., p. 133.
Digitized by VjOOQ IC
^0 ELKCTRIC SMELTING AND UEFINING.
molten glass* or in pure water. Sodium oxidises very rapidly
in the air, yet it may be melted without danger in a diy vessel
over an open flame, provided that it is not heated much above
its melting point. Once it is ignited in air that is dry and
free from carbonic acid, it burns with a yellow flame, evolves
much heat, and yields the peroxide, Na^O.^. It also combines
energetically, with the non-metals. It decomposes water even
at the ordinary temperature, forming sodium hydroxide, and
must, therefore, be stored in oxygen-free liquids, such as petro-
leum. As a metal which can decompose even water, it is
very violently attacked by acids, forming salts which are, almost
without exception, easily soluble in water. It is a powerful
reducing agent, decomposing compounds of the metals, and even
of many non-metals (COg, SiOg, BgOj, &c.).
After the experience gained in the electrolysis of fused
magnesium and lithium chlorides, attempts have naturally been
made to obtain sodium electrolytically from the readily-purified
sodium chloride in the molten condition, but, up to the present,
the practical results have not been entirely satisfactory. Up to
the year 1890 purely chemical methods for the reduction of
oxidised compounds of sodium were used in practice. Only
a very slight outline of these processes can be given here.
BEDUCTIGN PBOCESSES.
On the large scale, sodium was first f produced by distilling a
mixture of anhydrous carbonate (calcined soda) and carbon;
whilst until comparatively recently the bulk of the metal was
obtained by Castner's method,! which consisted in a reduction of
the hydroxide by an intimate mixture of iron and carbon,
thus : —
3 NaOH + Fe . Ca = 3 Na + Fe + CO + COj + 3H..
Netto § avoided the use of iron by causing fused caustic soda
to drop upon a layer of heated carbon contained in an upright
retort or reverberatory furnace.
EliECTBOIiYSIS.
Davy's Experiments. — The electrolytic decom{K>sition of
the hydroxides of potassium and sodium led, as is well known,
to the discovery of these metals. Davy|| tiixxB described the
apparatus which he employed for his research in this direction : —
* Zsigmondy ; see Zeitschr.fur Elektrochem., 1898, vol. iv., p. 546.
t Bninner, Schwtigger^s Joum,, vol. Ixxi., p. 201 ; and St. Claire Deville,
Ann. de Chem, et de Phys., 1852, vol. xliii., p. 5.
t United States Patent 342,897, June 1, 1886. [Eng. Pat. 7,395, 1886.]
§ German Patent 45,105 and 52,555. [Eng. Pat. 17,412, 1887.]
I) PhU. Trans, of 1808, pp. 1, 333 ; 1810, p. 10.
Digitized by LjOOQ IC
SODIUM. 31
*'By means of a stream of oxygene gas from a gasometer
applied to the flame of a spirit lamp, which was thrown on a
platina spoon containing potash, this alkali was kept for some
minutes in a strong red heat, and in a state of perfect fluidity.
The spoon was preserved in communication with the positive of
the hattery of t^e power of 100 of 6 inches, highly charged ; and
the connection from the negative side was made by a platina
wire."
The conduct of the process on a large scale was not possible
with an apparatus of the type used by Davy ; and Castner, in
1890, was the first to solve the apparatus problem, and so to
render Davy's process industrially practicable.
Davy had proposed another method, in which a platinum dish
<»ntaining mercury was connected up as the negative pole in a
concentrated solution of caustic potash or caustic soda, a galvanic
battery being used as the source of current. The alkali metal
thus separated by electrolysis formed an amalgam from which it
was to be recovered by distilling off the mercury. But the
method was not satisfactory, owing to the yield of metal bearing
too unfavourable a proportion to the expenditure of electrical
energy. Both proposals, however, contained ideas which have
since been repeatedly rediscovered — namely, the use of a vessel
of conducting material which should serve simultaneously as a
receptacle for the electrolyte and as one of the poles of the decomposim
tion cell ; and, further, the use of a liquid metallic cathode to
absorb the alkali meted during the electrolysis of its aqu^eous
solution.
ELECTROLYSIS OF SODIUM CHLORIDE.
Watt's Prooess. — The first suggestion for the manufacture of
alkali- and alkaline-earth metals originated with Charles Watt.*^
The patent specification ran as follows : —
" The second part of my invention consists of a mode of pre-
paring or obtaining the metals of the alkalis and alkaline earths
by the united action of electricity and heat. For performing
this part of my invention by the united action of electricity and
heat I employ a vessel of the form shown in Fig. 17, wluch is
made of iron or other suitable material capable of bearing a full
red heat. In this figure, A is the vessel, which should be at
least half an inch thick, and, if made of iron, previously to its
being used should be coated over its exterior with day or other
substance to preserve it from the action of the fire ; B, movable
head for the collection of the metals ; C, electrodes, with their
attachments, e ; D, flanges to support the vessel upon the furnace.
The covered compartment, F, \mng that in which it is intended
to eliminate the metals, is supplied with a carbon electrode, and
* Eng. Pat. 13,755, Sept. 26, 1851.
Digitized by VjOOQ IC
32
ELECTRIC SMELTING AND REFINING.
the uncovere<l compartment is supplied with a gold electrode ;
but I wish it to be understood that I do not restrict myself to
the particular form of apparatus, or to the material to be used
for electrodes. The vessel is filled with dry saline matter, so that
when it is in a state of fusion it shall reach the dotted line [the
level shown in the figure] ; the partition keeps the eliminated
substances from reacting upon each other, and also excludes the
air from the compartment in which the metal is eliminated, the
access of which would cause the metal to be oxidised. The
vessel is placed in a furnace where it can be subjected to the
action of a full red heat, and when the saline matter is in a state
of fusion contact is made between the decomposing vessel and
the apparatus supplying the electric current or currents, the
intensity of which should, at least, be equal to that which would
be supplied by ten cells of Daniell's battery arranged for in-
Fig. 17. — Watt's apparatus for the extraction of sodium.
tensity, but, of course, this depends upon the nature of the salt
which is being decomposed. The fused salt is maintained at that
temperature which will ensure the instantaneous volatilisation of
the metal as it is eliminated, and a propei* receiver (such a one
as is usually employed for the preparation of such metals will
answer) is connected air-tight with the narrow tube projecting
from the head. The metal is received and preserved in any con-
venient fluid hydrocarbon. The salts which I usually employ
are the chlorides, iodides, or bromides of the metals of the alkalis
or alkaline earths."
The salts which the inventor used may have been the haloid
compounds of the alkali- and alkaline-earth metals, as he
describes : but neither before nor since the publication of that
specification has it been possible to distil in an iron vessel the
Digitized by V^jOOQ IC
SODIUM.
33
metals calcium, barium, or strontium, of which the melting
points even approximate that of iron. Granted that he merely
omitted by accident to explain in his specification that the elec-
trodes were insulated not only from one another, but also from
the iron crucible, the inventor must have used a remarkably
refractory modification of gold if he succeeded in making from it
an electnxle which could be employed as the positive pole in
a bath of fused alkaline chloride, and thus withstand the action
of the nascent chlorine for a sufficiently long time to enable him
to accomplish the decomposition of the alkaline halogen salts
according to his ugual methods. The patent was applied for on
September 25, 1851, and granted on March 24, 1852.
It is well known that in the summer of 1851, Bunsen suc-
ceeded in decomposing fused magnesium chloride into magnesium
and chlorine by means of the electric current ; and this classical
research is described in vol. ii. (p. 137) of Liebig's Annalen for
1852. In outward form the apparatus used by Watt shows the
closest agreement with that employed by Bunsen, and the
difference between them lies in the fact that the latter, which
was based on the residts of Bunsen's investigation, fulfilled
its purpose in every respect, whilst the former was altogether
impracticable.
The first success in the electrolytic treatment of the chloride
of an alkali metal was obtained by Bunsen and Matthiessen
with lithium chloride in the year
1854. This result has already
been described (p. 22). Further
improvements by Matthiessen,
and afterwards by Linnemann,
related specially to potassium,
and will be referred to under that
head.
JablochkofPs Prooess. — In
spite of Bunsen's discoveries,
the decomposition of sodium
chloride appears to have involved
considerable difficulties, for until
the year 1882 this problem is
scarcely referred to in literature.
In that year, Jablochkoff* de-
vised the apparatus shown in
Fig. 18. Here the salt was
introduced through the charging funnel, D, into the clay crucible,
A, where it was melted and electrolysed. The electrodes, a and
b, were surrounded with the tubes, c and c^, for the removal of
chlorine and sodium vapour respectively.
The practical difficulties which arise in applying such an
• Dingler's polytechnisches JoiimcU, 1884, vol. ccli. , p. 422.
3
Fig. 18. — Jablochkoff's apparatus
for the extraction of sodium.
Digitized by
Google
34 BLECTRIC SMELTING AND REFINING.
arrangement to the electrolysis of alkali-metal chlorides are
unquestionably great. They lie chiefly in the designing of a
sufficiently durable apparatus. The tube for the removal of the
chlorine must not be made of metal ; and porcelain, especially in
close-grained pieces of the thickness required for the purpose, is
very easily broken at the high temperature employed. Then,
again, if the tube arranged to convey the metallic vapour to the
condenser be made of porcelain or any other material prepared
from clay, it will very soon become perforated by the action of
the alkali metals in the liquid or vaporous condition. But a
metal tube is equally unsuitable. Even if the electrode, 6, were
insulated from the surrounding tube, the metal sheath would be
of little value ; because, granted even that it might be possible,
by the use of sufficient heat, to maintain so high a temperature
that the separation of fluid metal — and therefore the formation
of a metallic contact between b and c^ — could be avoided, there
would necessarily be a deposition of sodium on the outside of the
tube throughout the process. This results from the fact that
insulated metallic substances placed in the electrolyte tend to
form intermediate electrolysing surfaces between the poles, the
side of the metal next the anode functioning as a cathode, and
that next the cathode acting as an anode. The consequence of
this is that intermediate reactions are produced, which cannot
but be prejudicial to the working of the process as a whole.
Under these circumstances the durability of the metallic sheath
is naturally of the shortest. Later modiflcations of the joriginal
construction, consisting in the arrangement of the electrodes and
of their separating partitions in the form of concentric tubes,
have failed to overcome the difficulty.
Hoepfher's Process. — In the year 1884, C. H. W. Hoepfner*
patented a process, of which the following is an account derived
from his specification: — "Sodium chloride is melted in a crucible,
the bottom of which is covered with a layer of copper or silver.
Instead of the latter, other heavy metals may be employed, with
the exception of mercury, which boils at too low a temperature.
The side-walls of the crucible must be made of some non-con-
ducting material. The layer of metal at the bottom of the
crucible serves as an anode, and it is, therefore, connected with
the corresponding pole of the galvanic battery, or of the dynamo,
by means of an iron or copper wire introduc^ either from below
or at one side. If the circuit be now completed by immersing
a carbon or metal cathode in the fused chloride from above, a
rapid deposition of metallic sodium at once sets in ; the sodium
bums in the presence of air, but if the latter be excluded, the
metal may be collected or distilled. While the light metal
separates at the upper surface, the chlorine attacks the metal
of the anode, forming a chloride which fuses at the high tern-
♦ German Patent 30,414.
Digitized by LjOOQ IC
SODIUM. 35
perature of the operation, so that by continuing the current
the anode undergoes electrolytic solution ; but on account of its
weight the chloride remains at the bottom of the crucible.''
Even after quite a short experimental trial it can be easily
ascertained that, with long-continued current, the separation
does not proceed as smoothly as it at first has the appearance
of doing, for the metallic chlorides do not remain as quietly at
the bottom of the vessel as the patent specification suggests.
With the density of current required for the deposition of
sodium, the copper, silver, or other metal dissolved at the
anode would, in a short time, be re-deposited at the cathode
in such quantities, and in so slightly coherent a condition, that,
in spite of every precaution, it would be washed off again by the
fused salt, and (especially in this form of apparatus) would be
returned to the anode. Not only would there be no chloring
evolution, which the invention aims at avoiding, but in a short
time there would be no deposition of alkali metal either.
Bogers' Prooess. — A. J. Rogers* has patented the apparatus
shown in Fig. 19, which is a longitudinal sectional elevation of
the whole apparatus along the line y y (Fig. 20), and in Fig. 20,
which is a plan of one portion of the apparatus along the line
XX (Fig. 19). A' shows the masonry, B the fire-grate. The
melting-pot, C, is provided with a cover, C, which may be
fastened down; on the middle of this cover is a funnel, c',
which closes automaticaUy by a ball-float valve, c, when the
crucible is sufficiently charged. The safety-valve, Cj, is pro-
vided to guard against risk from a sudden evolution of gas.
The decomposing-vessel, D, is divided into two compartments
by the wall, d, which is made of some porous material, such as
chamotte, and reaches nearly to the bottom of the chamber.
The tube, E, forms a connection between C and D, and a
cock, By serves to regulate the flow of melted material from C to
D. N is the negative, and P the positive electrode ; the wire
connections for the electrodes within the cells are insulated, and
at the same time protected from the attack of acid, gases, &c.,
by the chamotte tubes, n and /?, in which they are respectively
enclosed. The tightly-fitting cover, D', is connected by two
necks, F F', with separate condensers, and is fitted with four
closed doors at the points f f and /'/', shown in Fig. 20. The
separating partition, dy must be gas-tight even up to the top of
the roof, D'. Through the roof of the negative compartment is
passed the pipe, K, with the valve, ky by which hydrogen or
other reducing gas may be introduced. The vessel, D, is set
in thin masonry, and must be constructed of a non-conducting
and fire-resisting material, and should be about 1 foot wide,
3 feet long, and 4 feet deep. The decomposing-vessel bears a
strong resemblance to that of Watt, but it is not so faulty in
♦ United States Patent 296,367, April 8, 1884,
Digitized by LjOOQ IC
36
ELECTRIC SMELTING AND REFINING.
the details of its construction. Nevertheless, it is scarcely to
be expected that it could be advantageously used for the reduc-
tion of the alkali metals ; and the material of which it is made is
incompatible with great durability. The distillation of sodium
in clay vessels is an undertaking of very doubtful expediency.
The iron pre-heating vessel, C, filled at a bright red heat with
melted salt would, by its own weight, become unserviceable
Fig. 19.
Fig. 20.
Rogers' apparatus for the extraction of sodium.
after a few hours. To attach the tube, E, sufficiently firmly
to the clay vessel, D, and to maintain a tight joint when the
whole apparatus is charged with fused sodium chloride at a red
heat, would be no easy task. Rogers has, meanwhile, had an
opportunity of discovering the faults of the apparatus.
Omholt'8 Apparatus. — Omholt's plant* (Figs. 21 to 24) for
the continuous production of the light metals, appears also to^
. • German Patent 34,727, June 6, 1885.
Digitized by LjOOQ IC
SODIUM. 37
l)elong to that class of inventions which are first patented and
then, when opportunity offers, are put to the practi<^ test. The
patentee writ!^: — '^The apparatus shown in longitudinal section
(both in elevation and in plan), in Figs. 21 and 22, consists of
a reverberatory furnace, with the hearth divided into two
separate compartments by the partition, a a. In each of these
compartments are two half retorts, h and c, borne horizontally
and parallel to one another on supports, d d, of fire-resisting
material, placed side by side in such a way that the half retorts
are separated by a short space from the sole of the hearth.
The half retorts marked b b contain the negative electrodes, e e,
whilst those marked c c enclose the positive electrodes,//. The
halogen compound to be electrolys^ lies melted on the heartli
at such a height that the electrodes in the half retorts are com-
pletely immersed. The metal separating at the cathode, and the
halogen set free at the anode, collect witiiin their respective half
retorts, and are there protected from contact with the furnace-
gases, owing to the liquid seal which is foimed by the dipping
of the half retorts into the melted charge. The half retorts are
made of a fire-resisting and non-conducting substance, and are
coated on the interior with a lining of carbonaceous material.
The latter material consists for the most part of carbon (graphite
or wood-charcoal) ; the mixture used, for example, in the manu*
facture of plumbago crucibles being suitable. The admixed clay
serves only to give plasticity to the carbon.
"The electrodes are made of carbon or other resisting material.
They lie lengthwise within the half retorts, as shown in the illus-
tration, whilst at the one side (at g in Fig. 23) they penetrate
the masonry of the furnace, and there make the necessary con-
nections for the circulation of the electric current. To prevent
the separation of halogen or metal on the electrodes outside the
half retorts, the exposed portions between the faces of the retorts
and the brickwork of the furnace are insulated by a covering of
chamotte. By this construction, I [the patentee] allow for the
renewal of broken half retorts without damaging the electrodes
or altering their position.
"The half retorts, b 6, are connected with the fire-resisting fore-
chambers, kkj hy short tubes, h h, which should be as wide as
possible and which are best placed vertically ; these tubes, h A,
are made of some mixture containing carbon. The position of
the upper edges of the tubes, A A, determines the level of the
melted charge in each compartment of the hearth. The metal
separating at the electrodes, e c, collects under the half retorts, b 6,
on the upper surface of the fused salt, whence it passes, either as
liquid or as vapour, through the tubes, h h, into the fore-chambers,
kk, beneath, where it accumulates in vessels placed to receive it.
Each fore-chamber, k, is closed by a mouthpiece, n, similar to
those used in connection with coal-gas retorts, through which the
Digitized by LjOOQ IC
38
ELECTRIC SMELTING AND REFINING.
vessel placed beneath the pipe, A, may be withdrawn when it is
fully charged with metal, and replaced by an empty one. These
Fig. 21.— Omholt's sodium plant (longitudinal sectional elevation).
Fig. 22. — Omholt's sodium plant (plan)
Fig. 23.— Omholt's sodium plant (cross-section).
vessels are removed as quickly as possible from the fore-chambers,
and are allowed to cool either in a neutral gas or in such a way
Digitized by VjOO^ It!
SODIUM.
39
that air is excluded. A neutral gas is also passed into the fore-
chambers, k ky to the complete exclusion of air.
" The halogens produced in the half retorts, c c, are led away
after the manner of removing the metallic vapours just described,
but the upper margins of the upright tubes, //, project slightly
above the level of the fused salt in the furnace, in order to guard
against an overflow of the liquid into the fore-chambers, m wi.
The halogen is conveyed from the latter by means of a pipe con-
nection. The halogen compound may be charged into the hearths
continuously or intermittently, and either in the solid or in the
fused condition. To prevent the contamination of the melt by
ashes from the fuel or the like, the furnace may be constructed
like a muflle or muflle-fumace. The use of gas-firing is also to
Fig. 24. — ^Omholt's moditied sodium plant.
be recommended. To facilitate the exchange of worn-out or broken
half retorts, the furnace arch may be constructed with removable
chamotte blocks, o o. Instead of arranging the electrical con-
nections of the electrodes at the side, as in tig. 23, it is more
practicable to make them from below, as in Fig. 24, where the
electrode, «, is supposed to be the cathode ; the anodes, however,
are similarly connected up with the dynamo from below.
" To avoid the use of neutral gases, I [the patentee] effect the
removal of the light metal accumulating on the sunace of the
melted charge in the half retort, 6, in such a way that it passes
through the upright tube, h (Fig. 24), into a fore-chamber, k,
which is so far filled with inactive liquid («.^., a petroleum
product of high boiling point) that the curtain wall, S, of the
compartment) k, forms a liquid seal, preventing the access of the
outside atmosphere. The light metal flowing through h collects
Digitized by
Goog
vc
40
ELECTRIC SMELTING AND REFINING.
in a movable vessel, w, placed beneath, which may be drawn out
when full, i-aised through the entrance, u v, of the fore-chamber,
K, and replaced by another empty vessel."
So far the patent specification. No one who has had any
experience in the electrolytic separation of the metals of the
alkalis, earths, or alkaline-earths in vessels, which are placed in
direct contact with the furnace-gases, and which are made of such
porous material as are the half retorts described by Omholt, will
have a moment's hesitation as to the verdict upon this process.
The furnace-gases have, especially in the reverberatory furnace
just described, the most favourable opportunity of circulating in
intimate contact with the fused chloride, whilst they are only
separated from the vapour of sodium by a partition that is
extremely pervious to gases. The rate at which the metal is
.separated is reduced in proportion as the resistance of the electro-
lyte increases ; the rate at which the bath of chloride becomes
loaded with oxides and hydroxides being proportionately very
great. This latter circumstance is a result of the contact of
the hot furnace-gases with the charge, either directly or after
diffusion through the partition wall, and is in part due to the
oxidation of the metal and in part ix) the decomposition of the
chloride.
The diminishingly small yield of metal that must be predicted
for it, and the proportion of impurities in the bath, increasing
from minute to minute, and finally causing the complete stoppage
of the current, would alone suffice to condemn an apparatus of
this description. In passing un-
favourable judgment upon this
plant, no account has been taken
of the losses resulting from the
greater irregularities that must
occur, such as fiucture of retorts
and other evils.
Fischer's Process. — F. Fischer*
has described his experiments with
the arrangement shown in Fig. 25,
as follows : — ** An iron crucible is
divided into three compartments
by two transverse partitions, reach-
ing nearly to the bottom. The dry
alkaline chloride is introduced into
the first division in such quantity
that the level of the melted salt in the closed decomposition-
compartment may stand at about that of the line, a. The carbon
plate, c, which serves as anode, is either continued upwards
through the cover of the cell, or makes connection with the
electric generator by a protected metal wire. An iron plate, e,
* Wagner-Fischer'd Jahresbericht, 1886, p. 222.
Digitized by LjOOQ IC
Fig. 25.-^501161^8 apparatus.
SODIUM.
41
may serv'e as cathode, and ihe sodium vapour produced is carried
off by a side tube. The metal walls of the anode compartment
must be protected from the action of the chlorine as it is evolved
and conducted away ; but this difficulty would disappear if the
carbonate were used for the decomposition." The electrolytic
dissociation of the carbonate, however, offers very great diffi-
culties, in consequence of the low electrical conductivity of this
salt and the ease with which carbon dioxide is reduced by
sodium. The use of the chloride leads to the same troubles that
are experienced with Troost's apparatus (p. 22).
nomung and Kasemeyer's Plant.— The apparatus of
Homung and Kasemeyer* is distinctly retrogressive. This
arrangement is shown in Figs. 26
and 27: "In which A is a graphite
anode, serving also as a crucible
for melting the charge, and enclosed
within an iron sheath. Attached
by the usual devices to the rim of
the crucible is the ring-shaped
cover, P, made of a close-grained
porc^in, and provided with a
channel, leading to an off-take main,
C Within this cover is hung a
,M
Fig. 26. Fig. 27.
Homung and Kasemeyer's apparatus.
[tube-shaped] porcelain screen which is long enough to be im-
mersed to a sufficient depth in the electrolytic bath. Through
this screen the hollow cadiode, K, made of iron or other suitable
metal, is lowered into the space within the anode. The cathode
must be of such width that a space is left between it and the
tube, S. At the upper end of the cathode a compartment, e, is
formed, of which the under side makes a tight joint with the
cover, P, and the shell, S. This chamber is closed above, but
is provided with a number of perforations, A:, below, through
which free communication is made with the space between
the cathode and the tube, S. At the side of the chamber is a
tube, M.
* German Patent 46,334, Jan. 29, 1888.
Digitized by VjOOQ IC
42 RLBCTKIC BMELTINQ AKD BBFINIKO.
"The cathode is open at both ends; and through the upper
opening, the original chaise is introduced, with the aid of any of
the usual mechanical devices for such a purpose, and afterwards
chloride is added, to make good that which has been decomposed
during the electrolysis ; this is done in such a manner that the
cathode may always remain perfectly full of salt, which thus
forms a close cover to the apparatus at this place. The lengths
of the cathode and of the anode, respectively, are so chosen that
the weight of the chloride charge in the former may suffice to
maintain the level of the liquid in the anode-cell at such a height
that all communication is cut off between the two compartments
formed by the sleeve, S. The chlorine formed at the anode
and the sodium vapour produced at the cathode, during the
progress of the operation, are kept apart by the sleeve, S, the
former passing to the tube, C, the latter to M. In proportion to
the decomposition of the salt, fresh chloride sinks into the bath
from the cathode, and this is continuously replaced by fresh
additions above.'
The whole arrangement strikes one as being a most un-
fortunate combination of devices of which each separately has
been thoroughly serviceable amid its original suiTOundings.
For example, the use of a carbon crucible as a combined decom-
position vessel and anode has proved successful in the apparatus
which Bunsen* employed in 1854 for the reduction of chromium
(see Chromium), The employment of hollow electrodes through
which the solid electrolyte might be charged, dates back even
farther, for J. H. Johnson's application for provisional protec-
tion in England, described such an arrangement in 1853 (see
Aluminium^, The inventors have now discovered that the
sinking of the substance in the tube does not take place un-
interruptedly, and have introduced a rotating spiral within the
electrode in order to prevent stoppage. The separation of the
electrodes, and of the substances produced upon them, by means
of non-conducting partition walls, is a principle as old as electro-
chemical decomposition itself.
If the metal and halogen are to be separated in Homung
and Kasenieyer's apparatus, it is necessary to replace the short
separating tube by a cell closed at the bottom and provided
with perforations at the sides. But even then the life of the
iron crucible cover could not be very long, since the metal is
surrounded with the heated furnace gases without, whilst within
it is in contact with chlorine which has diffused through the
substance of the graphite crucible. Among the apparatus
previously described there are some which, although they have
a few faults, yet possess the advantage of comparative simplicity-
Grabau's Apparatus. — ^Although the ideas embodied in
some of these inventions are undoubtedly good, it is not pos-
* t^oiig, Ann,f 1854, vol. xci., p. 619.
Digitized by LjOOQ IC
SODIUM.
45
sible that the processes could be employed satisfactorily for
continuous work, nor could they compete with the chemical
methods employed for the reduction of the alkali metals from
their hydroxides or carbonates. That the source of the difficulty
lies in the technical application of the process is shown by the
account published by A. J. Rogers,* who was able to produce
2*5 to 3 kilogrammes [5-5 to 6*5 lbs.] of sodium from the chloride
per electrical horse-patver in 24 hours, provided that the apparatus
teas sufficiently durable to allow of uninterrupted work. The
Fig. 28. — Grabau's apparatus for the reduction of alkali metals.
elec.tro-chemical efficiency of the process was therefore very high,
but the apparatus employed was still imperfect. In the same
year that Rogers' paper was published, Grabau patented his
apparatus which was intended primarily for the electrolytic
extraction of alkali metals from their halogen compounds ; the
patent specification f was published on May 2, 1890. In this
arrangement the melting vessel, A (Fig. 28), is surrounded by
* Journal of the FraiUdin Inst,, 1889, vol. cxxviii., p. 486.
t German Patent 51,898, October 8, 1889. [English Patent 15,792,
October 8, 1889].
Digitized by LjOOQ IC
44 ELECTRIC SMELTING AND REFINING.
an air-bath, L, heated by the hot gases, which circulated in the
space, G, around it. Within this vessel are a bell-shaped
cathode-cell, B, made of porcelain or other suitable fire-resisting
material, and the carbon positive elect-rodes, C, which are
arranged around the latter. The whole is closed by a cover, D.
From the lower lip of the cell, B, the wall, m?, bulges out and
is continued upwardly and then over the level, N N, of the fused
electrolyte. In this way a space is provided between the wall of
the cell, p, and the outer jacket, w ; and as melted matter cannot
penetrate this space, the fused salt does not come into contact
with the outer surface of the wall, f», and it is thus impossible
that the cell-walls can permit any direct electrical connection
to be made between the melted matter within and that without
the cathode cell.
The electrode cell, B, is connected by a tight joint, with a
hollow iron cap, E, from which a side tube, a, passing over the
rim of the melting vessel, makes communication with the receiver,
S. The cap, E, forms the negative pole, and makes electrical
contact with the fused salt by means of the rod, n, placed within
the bell. The screw-plunger attached to the frame, H, is provided
for the purpose of removing any stoppages that may occur in the
tube, E. Since the alkali metals are specifically lighter than
their chlorides in the fused state, the fluid metal accumulating
in the cathode cell is forced upwards by the pressure of the
melted electrolyte, and flows oflf through the tube, a, as fast as it
is produced. The collection may, for example, be effected in a
petroleum filled vessel, S, in which the cylinder, M, can be
filled with a neutral gas admitted through the tube, c. The
chlorine evolved at the anodes during electrolysis escapes through
dj whilst the charge of salt is introduced into the melting vessel
through e.
In a supplementary patent dated September 19, 1890, Giabau *
recommends the use of a mixture of three chlorides (of sodium,
of potassium, and of an alkaline-earth) in equivalent molecular
proportions, the advantages thus gained being : — a greater fusi-
bility of the mixture, the prolongation of the life of the vessel,
and an improved yield of metal. The resulting sodium is said
to be nearly free from potassium and alkaline-earth metal.
Grabau approached with his apparatus very near to the solu-
tion of the problem of sodium-extraction by the electrolysis of
common salt ; indeed, even in this form an apparatus may, under
fortunate circumstances, have a life of several days, or even
* [English Patent 16,060, October 9, 1890. In this specification Grabau
proposes to employ equi-molecular proportions of potassium and sodium
chlorides, and to add 2 molecules of strontium chloride for every 6 mole-
cules of the mixed alkaline chlorides. Strontium chloride was preferred
to the calcium salt, because it is more easily obtained in the anhydrous
state ; but either (or barium chloride) may be used. The sodium produced
contains no strontium, but may retain 3 per cent, of potassium. — Trans.]
Digitized by V^OO^ It!
SODIUM.
45
weeks ; but the starting of the process would necessitate many
breakages, and every derangement of the plant that calls for a
suspension of the process might entail the loss of the whole
apparatus. The mixing of several salts to effect a reduction of
the melting point is not new, for Bunsen and Matthiessen have
frequently referred to the principle ; and the same may be said
of the observation that, under certain circumstances, the alkali
metal is alone separated in the electrolysis of mixtures of alkali
metal and alkaline-earth chlorides.
In the author's experiments, which he undertook rather
with the object of obtaining alkali metals than with that of
producing pure sodium, he started with the very fusible mixture
Fig. 29. — Borchera' apparatus for reducing the alkali metals.
of KCl + NaCl ; but he has demonstrated that even in this case,,
if the current-density be not excessive, and if fresh sodium
chloride be introduced as fast as this salt is decomposed, the
metallic sodium obtained by electrolysis will conteun only a
small proportion of potassium.
Borohers' Apparatus. — The apparatus used is shown in
Fig. 29. The melting vessel, G, had an opening surrounded by
the socket tube, M, and two other openings with tubular necks, R.
The double-socket porcelain tube, J, fitted into the socket, M,
and in this again fitted the iron socket tube, N. N was closed
above by an iron plug, through which was passed the iron
Digitized by V^jOOQ IC
46 BLBGTBIC SXBLTINO AND REFINING.
rod, K ; the latter served as cathode, and was therefore partly
immersed in the fused salt. The sodium separating upon the
lower part of the cathode floated upwards through the tube, N,
and collected there until it overflowed through the side tube, ti,
into a suitable vessel placed to receive it. The level of the
melt was, of course, maintained sufficiently high to enable the
overflow of the sodium to take place; and the liquid was
prevented from forcing its way through the flanges by means of
metallic lead.
Within the neck, R, the porcelain tube, C, was suspended,
and within this again was the carbon anode-rod. A, which was
held by the copper grip, H, resting upon the annular cover, D.
The chlorine was conducted away tlux>ugh the tube, e. A third
tube (not shown in Fig. 29) of the same height and width as R,
served for the introduction of the electrolyte. The apparatus
was designed for use with a current of 30 to 50 amperes, and
afibrded a yield of from 65 to 70 per cent, of the weight of metal
theoretically possible.
This apparatus has not proved to be very durable. Cast
iron, at a red heat> will not for long withstand the effects of
alkali-metal chlorides on the one side and of furnace gases on
the other. Of the porcelain components, the double-flanged
tube was found to be very readily broken, a fault which would
become more serious as the size of the plant was increased.
Beqiiirements for an Eleotrolytio Sodinm-Extraotion
Apparatus. — Both this apparatus and that devised by Grabau
possess the disadvantage that sodium is brought into direct
contact with porcelain, for there must inevitably be a loss of
sodium caused by the action of the metal upon aluminium
silicates at these points, whilst the porcelain itself is rapidly
corroded. The lessons that may be derived from the negative
results hitherto obtained may be summed up shortly in the
following rules for the construction of a suitable apparatus
for sodium extraction : —
1. A refractory metal only may be used as the material for the cathode —
preferably the better sorts of iron.
2. The alkali metal must be collected in, and conveyed from, the cathode
cell without coming into contact with any reducible substance.
.3. The walls of the cathode chamber may be made to serve also as
cathodes, but in that case they must not be in contact with the electro-
lyte on their outer surfaces.
4. The anodes must be made of carbon.
5. The anode compartment must allow of an easy escape for the halogen ;
and its walls must be made of some material that will withstand the action
of the halogens and haloid salts.
6. The walls of the anode compartment must not be in contact with the
Beparated metal
7. The electrolyte must not have any metallic object in it between the
poles or in the path of the current.
8. The whole apparatus must be made of a fire-resisting materiaL
Digitized by VotOOQIC
SODIUM.
47
These conditions were at that time fulfilled in only two forms
of apparatus; one was that introduced by Grabau in 1891, as an
improvement on his earlier invention. It differs from it only
in the arrangements of the cathode bell, so that this alone now
requires explanation.
Grabau'8 later Apparatus. —
The modem Grabau cathode cell
has a wider space than that origi-
nally provided between the bell
and its bent outer wall. In this
space is arranged a cooling tube
with connecting pipes, Z and A,
for the admission and removal re-
spectively of a cooling liquid. The
object of this is to maintain a thin
crust of solidified salt upon the
outer wall of the cell; and in order
that the crust may not become too
thick the cooling tubes within the
U-shaped jacket of the bell are
packed in a material having a low
conductance for heat. The tem-
perature of the fused salt within
the bell is maintained sufficiently
high to prevent solidification on
the inner walls, owing to the enor-
mous density of the current that
is here passing through the elec-
trolyte.
Borchers' Sodium-Extraction
Plant. — Another form of appar-
atus was described by the author*
in 1893. This apparatus is shown
in Figs. 31 and 32, which are
about one-eighth of the full size.
It consists of two chambers con-
nected together by a special joint.
One of these, the cathode chamber,
K, is of iron, whilst the other,
the anode compartment. A, is of
chamotte. Thus the anode, a, is
provided with a shield through which no chlorine can pass to
the cathode, and which cannot be attacked from without by the
used salt, or from within by the powerfully-reducing metal
obtained at the cathode. It is in this way possible to use tube-
shaped cathodes for the collection and removal of the metal,
* Borcbers, " Alkalimetalle," in the ZeiUehrift fUr angewandU ChemUj
Fig. 30.— Grabau's modified
cathode cell.
Digitized by VjOOQIC
48
ELECTRIC SMELTING AND REFINING.
the advantage of which is that the deposit can be produced
only upon the inner side, for the outer surface is nowhere in
contact with the electrolyte.
Great difficulties were encountered at first in making a good
joint between the vessels. It was accomplished, as diown in
Figs. 31 and 32, by means of a hollow metal water-cooled ring,
R, held together above and below by the clamps, Z. The only
really impermeable material that is available for such portions
Cooling W»ter: Escape). '|*-'CooliDg Water: Admission.
Figs. 31 and 32. — Borchers* sodium-extraction apparatus.
of an apparatus is the electrolyte itself solidified by cooling,
no other material being able to withstand for long the action
of the fluid salts. The breakage of the chamotte portion of
the apparatus is avoided by placing an asbestos washer between
the water-cooled ring and the flange of the anode compartment,
A, in order to minimise the difference of temperature between
the adjacent parts. A tube, C, is provided to carry off the
chlorine. A sieve, S, with a little asbestos resting upon it is
Digitized by V^jOO^ It^
SODIUM. 49
provided to receive the small crystals of added salt, which would
otherwise be apt to crack the anode vessel.
The cnrrentKlensity on the cathode, K, must amount to at
least 5,000 amperes per square metre [3-2 amperes per square
inch]. A smaller current^ensity at the anode surface would
necessarily lead to an economy of power, but it is scarcely to be
attained in an apparatus of the kind employed for the electrolysis
of fused lithium and magnesium salts. In this respect the new
arrangement offered distinct advantages, for, with the same
current-density at the cathode, that at the anode could be
reduced to any required degree, and the expenditure of power
must then show a corresponding diminution.
About the time when these results in the electrolytic treat-
ment of fused alkali-metal chlorides were obtained, Castner took
up Davy's process and rendered it practicable by the construction
of a very serviceable apparatus. But the description of this
apparatus, and of others which, like it, were designed for the
electrolysis of fused caustic soda, will be given after a short
account of the further experiments in the electrolysis of fused
chlorides.
Danckwardt's Apparatna. — Danckwardt, in the year 1898,
patented'*' for this purpose an apparatus in which, by the use of
water-jacketed or cooled walls, an electrolytic cell is obtained
which is capable of resisting the action either of fused alkali-
metal chlorides or of chlorine ; and to this extent the construc-
tion in question must be regarded as skilfully thought out. The
electrolytic furnace consists of two chambers, one, A, for the
anode and the other, B, for the cathode, built upon a masonry
foundation, a.
The walls of the chamber. A, are constructed of three iron
compartments, c, d, e, and the hollow partition-wall, /, which
separates A from B down to half the height of the chamber,
c and / rest on hollow supports. Cold water circulates through
the four compartments, c, d, e, /, gaining admission through c\
cP, c^, /*, and escaping through c\ li^, e^/^- The inner walls
of the compartments are lined with fire-resisting plates, g, to
protect them from the direct action of the heat. Four movable
carbon anodes, connected by A^ with the positive pole of the
dynamo, are passed through openings in the hollow supports; and
the spaces between the anodes are filled up with loam or clay.
One or more blastpipes, i, are passed horizontally through the
water compartment, e, and, by introducing air and gas (through
t^ and i*), serve to heat the charge of salt. A pipe, cP rf*, closed
by a clay stopper during the reducing operation, passes through
the chamber, rf, at the Attorn, and serves for the removal of the
molten chloride, which, when tapped off, collects in E. The
vertical walls of the compartment,/, are closed together, to allow
• U.S.A. Patent 607.506, July 19, 1898.
Digitized by ViiOOgle
50
ELECTKIC SMELTING AND REFINING.
Figs. 33, 34, 35, 36, and 37. — Danckwardt's sodium -ex traction apparatuB.
Digitized by V^jOOQIC
SODIUM. 51
for the admission of the fused salt from A to B. The floor, a,
of the anode chamber is cooled by means of the bent pipe, a\
A heavy chamotte plate, ^*, with manhole, J^, and gas escape-pipe,
j^, 8ei*ve8 as a cover.
The cathode compartment^ B, consists of an iron receptacle, k^
surroonded with masonty, a% to minimise the loss of heat, and
connected with the negative pole of the dynamo by means of the
leads, I V-. The cover of the receptacle is inclined slightly
towards the compartment, f, and, on the opposite side, ends in
the overflow pipe, k^. A series of iron plates, arranged vertically
on the bottom of B, serves to assist the distribution of the fused
material flowing from A into B, and also the flow of the chloride
into A at the time of tapping.
In order to set the furnace in operation, the anodes, A, are
drawn back, the cocks of the water inlets, a\ c^, d\ «^, f\ are
opened and the gas blast, t, lighted. The hot gases, after
heating A, flow off through B k^. When a sufficiently high
temperature has been attained in the two chambers, A B, the
gas blast is stopped, the carbons, h, are moved forward into A,
the space between them is well filled up with loam, common
salt is introduced through j^^ and the gas blast is again lit. The
}$alt melts, and in part flows into B; then fresh salt is added
until, when fused, the charge reaches almost up to the edge of
the overflow pipe, k^. The current-circuit is now closed, and
the tube, ^'*, is opened for the removal of the chlorine. The
alkali metal deposited on the cathode, m^ being of lower specific
gravity, rises through the fused charge, and flows through k^
into the receptacle, D. Salt is added to the charge from time
to time, as required, to take the place of that winch has been
clecomposed. At the end of the operation, the current is first
switched off, the clay plug is removed from the tube, rf*, until
the whole of the chloride has run off into E, and then the blast
is stopped.
Instead of melting the salt with the aid of a gas blast, the
first portion charged may be covered with a layer of charcoal,
which is then burned in a current of air. When this portion
is melted, alternate layers of salt and charcoal are charged
through y*. When the salt is completely melted, it remains
at the bottom while the charcoal floats on its surface, D^. In
other respects the process is the same as when gas-firing is used.
When it is desired to end the operation, charcoal alone is charged,
instead of salt and carbon, until the whole of the salt is used up.
The chlorine produced is not so impure as to be useless for
technical purposes. The proposal to use charcoal for heating
the apparatus is only mentioned as a possibility, without
having been actually tried. But no charcoal available is free
from ash,, and even the commercially pure sodium salt contains
sufficient impurities of itself without the introduction of others
with the ash of the fuel.
Digitized by V^jOOQ IC
52 ELECTRIC SMELTING AND RKFINING.
Eleotroly Bis of Fused Salt without External Firing. — An
inspection of the drawings of this furnace gives the impression
that it is an apparatus used for electrolysis, and wanned by
electrically-produced heat, and it is somewhat surprising that
until quite recently this method of working, which owes its
origin to the aluminium industry, has not also been applied ta
the electrolysis of common salt. A number of experiments in
this direction were tried in Borchers' laboratory during the year
1899. The first series of experiments was made by A. Fisdier,
and the following extract may be made from his account of
them: — *
"As in the experiments of Grabau and Borchers, a mixture
of salts (equivalent quantities of the chlorides of potassium and
sodium) was used, the mixture being more fusible than sodium
chloride alone, and the loss of metal increasing with the tem-
perature.
''Mutual Action of Sodium and Chlorine. — In order
to study the mutual action of sodium and chlorine after
separation at the electrodes, preliminary experiments were tried
in which no partition was placed between the electrodes. The
mixture of salts was stamped into a sheet-iron vessel, and an
iron rod cathode was then caused to approach a carbon anode
until an arc was formed between them, so that in a short time
sufficient salt melted to form a conducting mass between the
electrodes, which were then gradually separated until a well of
molten salt was formed (see Fig. 38). The surprising fact was
now observed, that the sodium, as soon as it had united inta
globules, whether of large or small size, remained for a long time
untouched, even in the midst of the stream of chlorine bubbles
flowing on the surface of the fused bath from the anode to the
cathode. The larger globules were somewhat flattened in shape ;
when intentionally brought close to the anode they remained
there for a notably long time. The low specific gravity of the
sodium, and the colour of the metal burning with a yellow glow
upon the surface of the bath, facilitated ready observation under
the conditions of the foregoing experiment, even in the case of
small globules of metal ; and it is due to this circumstance that
one, and that probably the most important, of the sources of loss
was readily discovered in this research. At the current-densities
necessary for this work a very brisk stream of small bubbles of
chlorine rise to the surface, and there unite to larger bubbles,
many of which remain unaltered for a comparatively long time.
It can readily be understood that the bubbles of chlorine in rising
through the bath are accompanied by a flow of liquid. The
consequence of this is that in the lower part of the bath there is
a flow set up towards the anode, whilst at the upper surface of
the fluid charge the chlorine bubbles with the liquid adhering to
* ZeiUichr, fur Elektrochem,, 1900, vol. vii., p. 349.
Digitized by V^jOOQ IC
SODIUM.
53
them tend to flow from the anode to the cathode, such circulation
being entirely independent of that produced by electrolytic
action. Thi» flow of liquid is so active that the particles of
Fig. 3S. — Diagram illustratinir behaviour of sodium and chlorine
in electrolysis of fused sodium chloride.
sodium, detached from the lower portion of the cathode, are
carried far in the direction of the anode, in spite of the up-
ward tendency that they possess in
consequence of their low specific
gravity. If the bath be observed
from above it appears as though the
sodium rises to a point situated be-
tween the anode and cathode, and
not to the cathode itself, and as
though from this point it is then
drawn to the cathode. The direction
of the flow of liquid and of the par-
ticles of sodium in the bath is shown
by dotted lines and arrows in Fig. 38.
In the bath used, the globules of
sodium rise to the surface at a dis-
tance of about from 6 to 10 cm. [2*4
to 4 ins.] from the cathode.
" Fischer's Modifioations of the
Depositing Apparatus. — ^The out-
come of this experiment was to sur-
round the cathode with a hood
dipping a little beneath the surface
of the fused charge, and wide enough
to catch the globules of sodium torn
off from the electrode. The hood first
used for this purpose is shown in
Figs. 39 and 40; it consisted of a
copper bell surrounded with a cooling
ring through which water was caused to circulate, a partition
placed between the inlet and the outlet for the water, causing
Figs. 39 and 40.— Hood em-
ployed by Fischer to sur-
round cathode in sodium
chloride electrolysis.
Digitized by VjOOQ IC
54
ELECTRIC SMELTING AND REFINING.
the water to flow through practically the whole ring. The
hood was attached by means of a sleeve to the iron rod
which formed the cathode. This arrangement was adopted
because it was found to be necessary before lowering the hood
to fuse so wide an area of bath that, when lowered, the bell
should dip into completely fluid salt, for if it is to prevent the
combustion of sodium at the surface of the bath its lower rim
must be so far immersed as to include a considerable proportion
of the surface area. The object of cooling the edge of the bell
was to produce a solidifled crust of salt on the rim, which thus
becomes insulated, so that the separation of sodium is restricted
to the cathode rod and also to any unprotected portions of the
inner surface of the hood. The sodium was tapped off by means
of a pipe, as shown in the arrangement illustrated by Fig. 41.
The cooling effect was found, however, to be too great, for the
salt soon solidified beneath the hood, and sodium then appeared
Fig. 41. — Fischer's sodium-chloride electrolysis apparatus.
on the side of the rim nearest to the anode, and, burning, e^ olved
so much heat, that the crust was melted away in patches, allow-
ing the separation of the sodium to take place at these places
without hindrance.
** Fischer hoped to overcome this difficulty by the use of a larger
bell. In order that all unnecessary parts of the apparatus might
be dispensed with, the walls of the pipe used for the removal of
the sodium were made of such thickness that the pipe could be
made to act as cathode also (compare Fig. 41). The hood or bell
was made oval in form, and was so placed over the cathode that
the position of the front rim ensured the inclusion of all globules
of sodium within the space covered by the bell, so far as could be
predicted from previous experiments. The front rim was also
sunk more deeply in the bath than was the hinder rim, the sole
duty of the latter being the exclusion of air. But, even with
this apparatus, sodium soon began to deposit on the outer side of
the cooling ring at the part opposed to the anode, and quickly
Digitized by VjOOQ IC
SODIUM. 55
increased in quantity in spite of the current of anode chlorine
which had free access to it at this point.
" It was proved by various observations that in the arrange-
ment hitherto adopted the copper bell acted as an intermediate
subsidiary electrode from the first ; for although a soNd crust of
salt was formed immediately the cooled ring was immersed, this
crust at the moment of immersion was so thin and so warm that
its electrolytic conductance was not sufficiently reduced to pre-
vent the metallic substance placed between the electrodes
from acting as an intermediate electrode when the current was
switched on. The rapid increase in the amount of sodium
formed on the rim of the bell, and the early solidification of the
fused material under the hood, were thus readily explained ; but
at the same time an indication was given of the way in whicb
these difficulties could be overcome. All metallic portions of the
apparatus that were not to be included in the system of con-
Fig. 42. — Fischer's modified sodium-chloride electrolytic apparatus.
ductors must be provided as far as possible with sufficient
insulation before immersion in the bath. Experiments in this
direction led to the following form of apparatus : — The melting
vessel was made of sheet-iron, as in the first experiments ; and
the cathode chamber was fqnned by a water-cooling arrangement
lined with marble. The pipe for tapping off the sodium was here
again used as cathode (see Fig. 42)."
But all these experiments led to the same result. The
solidification of the fused material commenced at the cathode
compartment. But the fractured surfaces of the mass, after the
whole charge had been allowed to cool, and had then been
removed from the apparatus, showed that this solidification
always happened in a markedly characteristic way, whether the
cooling had been considerable or only slight in extent. A layer
of black-coloured material permeated with fine particles of
sodium extended from the upper metallic surface of the portion
Digitized by VjOOQIC
56
ELECTRIC SMELTING AND REFINING.
which had been in the cathode compartment to the portion under
the bridge separating the anode and cathode compartments, just
as if a viscous material had flowed from the one to the other.
This layer ended in the anode compartment, where it was
rounded off in a parabolic curve. Evidently the heat evolution
at the cathode was retarded, owing to the rapid increase in the
surface of the cathode caused by the separation of sodium ; the
fused mass became first viscous, then pasty, and finally solid.
During the period in which the mass was viscous finely-divided
sodium was doubtless entangled in the pasty mass, and thus
maintained the conductance of the latter, even when the charge
in the cathode compartment had solidified. In this way t£e
cathode was extended into the anode compartment, and metal
was deposited on the side of the separating bridge facing the
anode ; this metal would then obviously be for the most part lost
Probably this difficulty could have been overcome by a timely
increase in the current-density, but, unfortunately, at the time of
these experiments the strongest current available in the Aachen
laboratories was 240 am-
peres. On the other hand,
in view of the object of
the research, which was
the construction of an
apparatus and evolution
of a process for use on
an industrial scale, and
for long runs, the sub-
stitution of an apparatus
working on a smaller
scale would have intro-
duced difficulties the re-
moval of which would
have involved an unnecessary loss of time and money. The
question whether this form of apparatus last described may not,
after all, be satisfactory if higher current-densities be used,
must therefore be left unanswered for the present.
Meanwhile, however, experiments were made by other workers
to design an apparatus suitable for use with the current-density
available. It appeared to be pre-eminently necessary that the
apparatus should be so constructed that the sodium should be
removed as fast as it was formed, and before it had time by
accumulation to increase the cathode ai-ea appreciably. This end
was gained by the use of the form of cathode shown in Fig. 43,
made in the fashion of a spiral staircase, but with an inclined
plane in place of steps. In the figure, for the sake of clear-
ness, the angle of the incline is somewhat greater than was used
in practice. When this electrode is slowly rotated during elec-
trolysis in the direction of the arrow (Fig. 43), the sodium, as it
Fig. 43. — Special form of cathode for
sodium collection.
Digitized by LjOOQIC
SODIUM.
67
separates, flows up the inclined plane (which becomes narrower
in the upper part) and collects at the highest point, where it
finds its way through an aperture into the space within the tube,
from which, in any apparatus of large size (compare Fig. 44), it
could doubtless be removed by means of a ladle, although iliis
was naturally impossible in an apparatus of the small size used
for experimental purposes.
The form of apparatus shown in Fig. 44 would be used for
work on a large scale. It is arranged. for driving by means of a
pulley, through which the current connections are made with the
cathode. The pulley is mounted on ball-bearings in such a way
that the electrode may be rapidly withdrawn from the apparatus
in the event of any accident or disablement of the plant. The
collecting tube may with advantage be left open below to facili-
tate the liquation of any of the fused electrolyte which may be
carried up the spiral witib the metal.
Fig. 44. — Arrangement of cathode shown in Fig. 43.
Another form of apparatus used by the author for experimental
work in the Aachen laboratories, and adaptable for use with
either light or heavy metals, is shown in Figs. 45 and 46. The
cathode is here placed in the centre of a carbon crucible, which
serves as anode. For use with metals which are only deposited
at high current-densities it is specially important that the anode
area shall be as great as possible, since in that way there will be
an economy of energy, owing to the minimising of the i*esi8tance
to be overcome, whilst at the same time there will be a reduction
in the flow of material which, as shown by A. Fischer's experi-
ments (see p. 52), tends to carry notable quantities of small
metallic particles from the cathode to the anode. The globules
of metal of low specific gravity rising from the central cathode
may be retained in a vessel suspended in the bath from above.
Digitized by VjOOQIC
58
ELECTRIC SMELTING AND RRFININO.
from which they can be ladled out as in the case of Castner's
apparatus. But ia this case cathodes may be suspended from
above within the electrolytic vessel.
Formation of Sodium Alloys. — The above described diffi-
culties in the electrolysis of sodium chloride early directed the
attention of electro-chemists to the use of a fluid metal with
which the sodium was capable of alloying as it separated out.
Fig. 45. — Borcliers' electric furnace for experimental work.
Fig. 46. — Borchers' electric furnace for experimental work.
Among such alloys of sodium only those with mercury have
found a practical application in the chemical industry. Mercury,
however, at the ordinary temperature, can dissolve only a small
proportion of sodium without losing its fluidity ; whilst, on the
other hand, its boiling point is so low that it has only been used
in practice as the aithode in the electrolysis of solutions of salts.
In this case the amalgam is merely an intermediate product
Digitized by VjOOQIC
SODIUM. 59
which becomes decomposed again as fast as it is formed. Taking
into consideration the fact that such a use of mercury does not
produce a final product in metallurgical works, but only one that
is employed in the manufacture of caustic potash or soda, there
appears to be no need to introduce into these pages an account
of the many forms of apparatus proposed with this end in view.
Preparation of Iiead-Sodium and Tin-Sodium Alloys. —
Lead and tin have been tried experimentally as fluid cathode
metals for the electrolysis of fused alkali salts; but no appa-
ratus or process for this purpose has yet had any permanent
application.
Rogers* has given an account of the first experiment in thin
direction, as follows : — " During the last three years (1886-1889)
I have experimented on the reduction of sodium chloride, using
molten negative electrodes, and especially lead. Lead, tin,
cadmium, and antimony all readily alloy with sodium, a large
part of which can be recovered from the alloys by distillation in
an iron crucible. They can be heated to a higher temperature
than pure sodium in acid crucibles without the sodium attacking
the crucible. In the following experiments a dynamo machine
was used to supply the current. In one experiment a current
averaging 77 amperes and 33 volts was passed through molten
sodium chloride, contained in two crucibles arranged in series,
for two hours. Each contained 30 lbs. of salt ; in the first was
put 104 grms. of tin, in the second 470 grms. of lead, each serving
as cathode, connection being made through the bottom of the
crucible. A carbon rod was used for the anode. When at the
end of two hours the carbons were removed, and the crucibles
cooled and broken open, the lead was found to contain 96 grms.
of sodium, or 17 per cent. There were about 90 grms. of sodium
found in the tin alloy, or between 45 and 50 per cent."
The electromotive force used was unusually high, but in the
absence of any indications in the paper as to the size of the
apparatus, it is not possible to form any idea of the current-
density employed. This account, however, suffices to prove that
a practical method for the electrolytic production of lead-sodium
or tin-sodium alloys had thus been found. In another paper,
Rogerst expresses himself somewhat optimistically as to the
consumption of power in the extraction of sodium. He expects
to obtain from the fused chloride 5| to 6J lbs. of sodium
per E.H.P. per 24 hours. According to the results already
described, such a yield is scarcely to be hoped for; he, how-
ever, adds the special restriction, "provided the apparatus in
sufficiently durable to permit uninterrupted work during the
time."
•Proceedings of the }ViMOjmn Nat. Ifif^. Soc., 1889 (from Richards'
Alnminimn.
+ J'oumal qfthe Franklin Inat,, 1889, vol. cxxviii., p. 486.
Digitized by LjOOQ IC
€0
ELECTRIC SMELTING AND REFINING.
The (iiiiiculty of devising suitable apparatus, which had thus
again become a hindrance to progress, led Yautin to put forward
some proposed improvements. His first patents* do not call for
mention, but his latest specificalionf must be referred to, because
it describes as new an apparatus^ which had been patented in
England so far back as 1844. Napier, who was the original
inventor, proposed to separate metals from fused substances,
using as cathode a crucible of some conducting material, which
was coated with a non-conductor (slag) internally down to the
bottom. This description accurately describes the latest form
of apparatus described by Vautin, and illustrated in Fig. 47.
But apart from the absence of any originality in Vautin's
aiTangement, this apparatus cannot be durable in character
when used for the work for which it is described in the
patent specification. There is no
material known that could survive
continuous use, even for a few days,
as an insulator in contact with a
portion of the surface of an exter-
nally fired vessel, the vessel being
employed as cathode in a fused mix-
ture of the haloid salts of the alkali
metal, earth, or alkali-earth metals.
It may be argued that the lining
material Cmagnesia) used by Vautin
is cheap ; but even if it cost nothing
it would still have to be rejected,
because it is worthless for use in any
continuous operation, and continuity
is essential to the profitable appli-
cation of heat- or electric-energy.
Borchers' Plant for Sodium Alloys.— The earliest appa-
i*atus which the author applied to the production of a sodium-
lead, or other readily fusible, alloy, is shown in Fig. 48. The
iron melting vessel, K, consisted of a short cylinder terminating
in a hollow cone below. The inner wall of the conical portion
was provided with projections, which formed terraced grooves,
one above another. The uppermost (and deepest) groove served
for the reception and melting of the lead, which was introduced
into the apparatus through one or more funnels. The i-emaining
grooves were intended to intercept the lead which flowed over
them, so as to renew its surface as often as possible, and to
enable it to take up a greater proportion of sodium. For the
electrolyte, one of the mixtures of salts already given was to be
* English Patent 13)568, 1893. (Cf. ZeitHchrift fiir EteXirUechnik und
Ekktrochemie, 1894, vol. i., p. 139.
t English Patent 20,404, 1893.
X English Patent 10,362, 1844; and 684, 1^5.
Digitized by VjOOQ IC
Fig. 47.— Van tin's apparatus.
SODIUM.
61
used in the melted condition. The carbon anodes, A, with their
tube-shaped porcelain shields, C, were hung from the chamotte
cover, D. As in the analogous apparatus described previously^
the carbon rods were supported on the covers, d, by the clamps,
V, which also served to make the necessary electrical connections
with the positive leads, P; the return leads, N, being connected
up with the crucible itself. The tubes, R, carried off the chlorine
generated during electrolysis, and the melted alloy which accumu*
lated in the bottom of the cone overflowed by the pipe, X. The
Fig. 48. — Borchers' apparatus for the production of sodium alloys.
loss of salt which was decomposed during the progress of the
operation had, of course, to be made good. In order to fuse
the charge, and to keep it in a fluid condition, the crucible was
suspended by the flange, F, in a heating chamber which, in the
form here shown, had been found to be economical of fuel. A
system of side-firing was adopted ; the heated products of com-
bustion entering through the flue, H, passed' into the heating
chamber, and, being deflected upwards by the circular chamotte
baffle, W, returned through the annular space between W and
the furnace walls, M, finding an exit by the flue, Z. In order
Digitized by VjOOQIC
m
ELECTRIC SMELTING AND HEFINING.
to catch any matter that should leak from a defective pot, a
iiollecting channel, 8, was built in front of the fire-bridge, and
from this the matenal could be run off at will. For small experi-
mental installations, a large Fletcher^s gas burner sufficed.
This apparatus has since been improved by the removal of the
porcelain tubes, which are very liable to breakage and other
damage. With this object the upper part of the melting ve^ssel
has been raised and surrounded with a cooling ring (Fig. 49).
The inner wall of the vessel thus becomes covered with a
layer of solidified salt, which protects it from the action of
the chlorine evolved on the surface of the carbon electixxie.
Fig. 49. — Borchers* improved apparatus for the production of
Bodium alloys.
The lead may be introduced into the uppermost gi'oove of the
alloying vessel, either after the manner formerly described or
(as here) from a separate melting pot, E, placed above the firing
chamber. The alloying cone is connected by a pipe, leading
from the bottom, to a side reservoir, B, which is heated by the
waste flue gases. Hence the alloy may be removed for use or
for test, as desired. The anode may be made either of one thick
carbon rod, or of several smaller rods.
Digitized by VjOOQIC
SODIUM.
63
Fig. 49 shows the arrangement of masonry recommended for
large installations. A plant of this kind, fifteen times the
actual size of the above illustration, is adapted to a current
of 300 amperes, which corresponds to a current-density of about
5000 amperes per sq. metre [465 amp. per sq. ft.] of cathode
surface. The electromotive force required may be only 6 to 8
volts. If a lead alloy containing not more than 10 per cent, of
sodium be required, the electromotive force during the whole
process need never exceed 8 volts.
The cost of producing sodium would have been very consider-
ably reduced by the use of this apparatus, especially as the cost
Pig. 60. Fig. 61.
Apparatus for the produotion of lead-sodium alloys.
of the plant would have been less, and its durability greater,
than in the case of the corresponding plant required for the pro-
duction of pure sodium, if it had not been for an unfortunate
circumstance which led to unexpected losses. The sodium alloys
of lower specific gravity were found to float on the surface of the
stream of lead flowing beneath them, so that alloys containing
from 8 to 10 per cent, of sodium alternated with others contain-
ing scarcely 1 per cent. Under these circumstances there was a
heavy loss of sodium by the re-dissolving of the metal on the
Digitized by LjOOQ IC
64
ELECTRIC SMELTIKO AND REFINING.
BurfJAce of the saper-saturated alloy. An apparatus designed to
overcome the difficulty of the irregularity in composition of the
alloy is shown in Figs. 50 and 51 ; but it has not been brought
into use because other methods have taken the place of the pro-
cess for which the lead-sodium alloys would have been useful,
and so the interest in the question has disappeared.
An apparatus admirably adapted to the production of lead-
sodium alloys has been described by Hulin.'^ But from his
experiments in the direction of obtaining caustic soda by the
treatment of the resulting alloys with steam, it may be assumed
that there is no very great demand for the alloys in question.
SEPARATION OF SODIUM FROM THE HYDROXIDE.
The Castner Prooess. — As above mentioned, Davy's process
for the electrolysis of caustic soda has been rendered practicable
by the construction of apparatus introduced by Castner, f whose
process is now universally adopted.
Fig. 52. — Castner's electrolytic sodium-extraction plant.
The apparatus (Fig. 52) consists of an iron melting-pot, A,
about 450 mm. [1 ft. 6 ins.] wide and 600 mm. [2 ft. high], into
which the cathode, H, is inserted through the bottom. To hold
the cathode in position and to ensure a good joint, the lower and
somewhat narrowed part of the vessel, with the tube, B, 80 mm.
[3| ins.] wide and 800 mm. [2 ft. 8 ins.] long, which is attached to
it, are both filled with caustic alkali before starting the electro-
lysis, and this in a short time solidifies. The bath of caustic
♦ Zeitschr.fSr Mektrochem,, and Jahrlmchfur Elektroekem,, 1894-1897.
i- German Patent 53,121 ; and English Patent 13,356 of 1890.
Digitized by
Google
SODIUM.
65
alkali, E, is kept in the fused condition by a gas flame, G ; and
in this bath are immersed the anodes, F, which (for this electro-
lyte) may be made of metal. They are suspended from the cover,
and are separated from the cathoide by a cylindrical diaphragm
of wire mesh, M. Above M is placed the collecting tube, 0, by
which the metal (D) and hydrogen are kept separated from the
liberated oxygen, which finds an escape from the anode compart-
ment by the opening P, at one side of the cover. The tube, C,
is closed by a lid, N, which rests sufficiently loosely upon its
support to permit the escape of hydrogen. For the removal of
the liquid sodium Oastner uses a perforated ladle, which is
able to retain the metal by reason of the high surface-tension
of the latter, while the caustic soda drains away through the
perforations. The various parts of the apparatus are separ-
ated by asbestos card as shown at S. I and L are the dynamo
leads.
Becker's Apparatus. — Oastner's apparatus was afterwards
modified by Becker,* as shown in Figs. 53 to 55. A wide tube, a,
Figs. 53 to 55. — ^Becker's modification of Castner's sodiom
extraction apparatus.
la inserted in the bottom of a metal vessel, A, and through this
tube is passed the rod, by to the upper end of which is attached
the usual cathode, B. The lower end of the tube is closed with
a ring, a\ made of lava, porcelain, refractory stone, or the like.
The rod, &, passes through the middle of this ring. The tube, a,
is surrounded with a double-walled water-jacket or other cooling
* German Patent 104,955, Jan. 21, 1899
5
Digitized by LjOOQ IC
66 ELECTRIC SMEI/riNO AND RBVININO.
deyioe, provided with the object of causing the electrolyte
within it to become pasty, or, if possible, to solidify, so that
none may escape at the joint between the tube and the closing
ring, a\ The cathode, B, is of metal, or, if necessary, of retort
carbon, and is made somewhat conical in shape, to facilitate the
escape upwards of the metallic globules forming upon it ; but it
can be of any other shape that fulfils this object. Fig. 54, for
example, illustrates a special form of cathode, consisting of a
number of rectangular, square, or round bars, b\ cast on to the
metal support^ 6^, which is attached to the rod, 6. The object
of this device is to increase considerably the superficial area of
the cathode, and so to concentrate the deposited metal in a small
space. The anode, C, is ring-shaped, and completely surrounds
the cathode ; it may be made in one piece or in several, and is
of retort carbon or of metal, according to the nature of the elec-
trolyte with which it is to be used ; it is suspended by one or
more rods, c, which serve also as conductors. These supporting
rods are so fastened to the outer surface or other part of the
anodes that they are separated from the cathode by a distance
greater than that between the anode itself and the cathode.
The anode must not reach to the bottom of the vessel, A, and its
height must not exceed that of the cathode.
A metal cone, D, insulated from the apparatus, is suspended
above the cathode, and is intended to collect the globules of
metal floating to the surface of the electrolyte. Around the
edge of the cone is an upright, or nearly upright, rim, d, whilst
in the middle there is a rising tube of fairly large diameter, and
with thick walls. This tube, 6, is closed above either by a thick
hinged cover or otherwise, and is provided with an exit pipe, /,
inclined slightly downwards, and passing outwards through the
wall of the containing vessel, A, being insulated from it by the
ring,/^, of asbestos, porcelain, or other suitable substance. The
diameter of the metal cone, D, must be somewhat greater than
that of the cathode, B, and less than that of the anode, C, so
that ail metallic globules which become detached from the
cathode pass into the conical receiver, whilst the gases generated
at the anode escape around the outside. The conical collector,
D, must not be immersed in the bath to a greater depth than
the height of the rim, d, so that the electrolyte may not cover
the cone. The upper surface is therefore always exposed freely
to the air, and, in consequence, any excessive rise of temperature
is avoided. If the electrolyte in the vessel should become too
hot> the cone, D, may be cooled either by directing a current of
cool air upon it, or by allowing drops of water to fall upon the
surface, where they instantly evaporate. If preferred, some
other system of cooling may be used.
Fig. 55 shows, by way of exasuple, a conical receiver, the
conical portion of which is made with double walls, through
Digitized by V^OO^ It!
SODIUM. 67
which is passed a current of a gaseous or liquid cooling-medium
— the effect of the cooling action being the reduction of the
temperature to such a point that the metal produced does not
distil off, and, on flowing through the tube, /, is not oxidised
outside.
The receiving cone is connected to the negative conductor, g^
hj means of a resistance so calculated that only a very small
portion of the main current passes through it. In this way the
cone becomes a supplementary cathode, and the metal liberated
at the principal cathode becomes negative again as it rises in the
collector, whilst it flows along in contact with the inner surface
of the cone, and in the tube which conveys it out of the
apparatus. The space between the main cathode, B, and the
supplementary cathode, D, is traversed by the metal so quickly
that the sodium cannot again become dissolved in the electro-
lyte ; and as soon as the latter makes contact with the cone, D,
it forms, as it were, a part of the negative electrode, and cannot
be attacked by the bath. The form of the cone shown is only
given by way of example ; other shapes may obviously be substi-
tuted for it.
If the anode gases are to be collected, the whole apparatus is
fitted with a cover provided with a vent pipe.
The vessel, A, is supported on a foundation of masonry or on
a casting. Within it the electrolyte is maintained in a molten
condition by means of heat generated from the current passing
through the bath ; but, as only the middle portion is thus kept
fluid, a long life is secured to the apparatus as a whole. Since
the metal flowing from the tube, j^ is cooled to a moderate
temperature by one of the methods above described, it is possible
to receive it in a mould of any required shape placed beneath
the discharge pipe, provided that the metal is not of too oxidis-
able a nature. If, however, the metal is readily oxidised, a
mould may be used which can be closed nearly gas-tight, and
which is provided above with an opening through which the end
of the tube,y ^, is passed.
Bathenau and Suther's Frooess. — Rathenau and Suther ^
have observed that the light metals, especially sodium, adhere
best to electrodes with convex or conical tops, when dipped but
a short distance into the bath. In this case the metal is not
detached even by gases evolved at the same electrode, so that
diaphragms are not required for the electrolysis of fused caustic
soda. The sodium separated may then be ladled from the
cathodes into funnel-shaped vessels, where it is allowed to
separate by liquation from any soda entangled in it. The soda
IB then run off through a valve at the bottom of this vessel, and,
finally, the sodium itself is tapped into moulds through the
same outlet.
* German Patent 06,672 ; and English Patent 21,027 of 1890.
Digitized by LjOOQ IC
6d ELECTRIC SMELTIKO AND REFINING.
With this process it is possible to work with cathode carrent-
densities up to 10 amperes per sq. cm. [0*93 amp. per sq. ft.].
The other results obtained are kept secret.
SEPARATION FROM SODIUM NITRATE (CHILI
SALTPETRE).
Darling's Frooess. — The results obtained from the electro*
lysis of the readily-fusible sodium hydroxide led Darling to
experiment in the direction of obtaining sodium from Chili
saltpetre, with the recovery of nitric acid. Whilst caustic
soda melts at 320% the nitrate melts at 313° C. The diffi-
culties attendant on the separation of a metal like sodium from
so powerful an oxidising agent as sodium nitrate, especially at
a temperature of over 300* C, may have given rise to doubts
as to the practicability of the process when the patent specifi-
cation was first published, but Darling appears, nevertheless,
to have been successful in putting his invention into practical
shape. According to the account published by him,* an experi-
mental plant, consisting of twelve furnaces capable of decom-
posing 300 to 400 kilogrammes [6 to 8 cwts.] of sodium nitrate
in all per diem, has been in use since 1901 in the well-known
chemical works of Harrison Bros., in Philadelphia, U.S.A.
The form of electrolytic apparatus first adopted by himt was
not sufficiently durable, but led to the determination of the
most favourable conditions for the separation of the sodium —
namely, that the sodium shall not come into contact with the
molten nitrate. This, evidently, could only be done by placing
a refractory and at the same time porous diaphragm between
the anodes and cathodes. After a long series of fruitless experi-
ments, the following process was adopted for the preparation of
the diaphragm:— Magnesia, which had been fused in the electric
furnace, was crushed and pressed into the hollow space between
the walls of a double- walled vessel, made of perforated steel
plates. This vessel was 760 mm. [30 inches! in height, and
400 mm. [15| inches] in diameter, and, after the magnesia had
been pressed into place, the thickness of the walls was 100 mm.
[4 inches], so that the internal dimensions of the cylinder were
660 mm. [26 inches] in height and 200 mm. flf inches] in
diameter. Otherwise the construction of the furnace was of
the simplest. A cast-iron melting vessel, set in a simple firing
chamber, served as anode; a layer, 150 mm. [6 inches] thick,
of refractory insulating material was laid on the bottom of the
vessel, and the above^escribed diaphragm rested on the middle
♦ Joum, Franklin Inst,, 1902, vol. cliii., p. 65.
tU.SA. Patent 517,001, March 20, 1894; English Patent 5,808,
March 20, 1894.
Digitized by LjOOQ IC
SODIUM.
69
of this layer, leaving a free space of about 75 mm. [3 inches]
between the melting vessel and the diaphragm. This space
was filled with sodium nitrate, whilst the diaphragm received
a charge of sodium hydroxide. A short length of 4-inch
wrought -iron tube senred as cathode. According to the
author's patent specifications of the years 1899-1900, an iron
cylinder was also suspended between the melting vessel and
diaphragm to serve as the true anode; but this cylinder appears,
from the modem descriptions, to have been found unnecessary*
It is, however, shown in the accompanying Fig. 56.
Fig. 56. — Darling's apparatus for the extraction of sodium from
sodium nitrate.
The electrolyte is heated moderately, and, during electrolysis,
yields sodium at the cathode and nitrogen dioxide and oxygen
at the anode. The sodium is ladled out of the vessel, and the
anode gases are collected and converted into nitric acid.
I The iron sheathing of the porous diaphragm obviously susts as
an intermediate electrode in the circuit ; in thus taking part in
the electrolytic reactions it would clearly suffer in respect of
durability. Darling obviates this difficulty by diverting, through
the iron sheath of the diaphragm, about 0-05 of the current
Digitized by LjOOQIC
70 ELECTRIC SMELTING AND REFINING.
passing throagh the whole cell, in such a way that the iron
is charged positively throughout the whole of the run.
The electric fusion of the magnesia for the diaphragm was so
costly that Darling has finally adopted, as a cheaper substitute,
a mixture of finely-ground clay, burnt magnesite, and Portland
cement. The cement produced by casting this mixture, after
the addition of sufficient water, into the hollow space within
the iron walls of the cell gave, after it had set, a very lasting
and useful diaphragm.
Each furnace takes a current of 400 amperes and an average
pressure of 15 volts. External heating is only necessary in
preparing a furnace for work and during the time of changing
the diaphragms, which now last from 425 to 450 hours. The
heat produced electrically in the furnace suffices to keep the
charge in the right condition during the run.
USES OF SODIUM.
Among the older uses of sodium may be mentioned the pro-
duction of chemically pure sodium hydroxide, the reduction of
organic substances in the aniline colour industry, and the reduc-
tion of compounds of the rarer or more difficultly -reducible
elements. Its use in the manufacture of aluminium, for which
at one time considerable quantities were required, has quite
ceased since the introduction of the H^roult process. Latterly,
however, there has been a large demand for sodium for the
manufacture of sodium peroxide and of the double cyanide of
potassium and sodium. Of these compounds the former is
already widely used as a substitute for barium and hydrogen
peroxides; whilst the latter, prepared in accordance with
Erlenmeyer's reaction, by the fusion of potassium ferro-cyanide
with sodium, K^Fe^Cj^ + Na^ = Fcg + (4KCy + 2NaCy), ia
employed in the extraction of gold.
Digitized by VjOOQ IC
P0TAS8TUM. 71
CHAPTER IV.
POTASSIUM.
Ooourrenoe in Kature. — In nature it is found only in the
form of salts; as halogen salts, in sylvioe, KCl, and carnallite,
KCl.MgOlj.6H2O; as sulphates, in the alums, K2S04.Al2(SO.)3.
24H^O ; as silicate, in the felspars, K2Al2(Si04)2, and in mica
and numerous other minerals and products of weathered
minerals. It appears to play an important part in the life
processes of plants and animals ; and both vegetable and animal
residues, such as wood- ashes, the ashes of beet-molasses, and
the suint from wool- washing are valuable raw materials in
the potash industry.
Properties of the Metal. — Potassium (K'; atomic weight »
39, specific gravity = 0*865) is like sodium, in that it is white
and lustrous when freshly cut, but it is softer than that metal,
and both its fusing point and its boiling point are lower. It
fuses at 62"" C, and vaporises between 700^ and 750*" C, the
vapour having a green colour.
Its chemical properties closely resemble those of sodium, but
it is usually more violent in its action, and this is especially the
case in its reaction with water.
Eztraotion Prooesses. — So also, in the same way, the
methods of extracting this metal are quite analogous to those in
use in sodium manufacture, and it may therefore sufiice to refer
to almost everything that has been written in the preceding
chapter. It should be mentioned, however, that in the direct
reduction of potassium carbonate or hydroxide by carbonaceous
material a black porous compound of potassium with carbonic
oxide is often found in the retorts and receivers, and this, by
reason of its instability, has been known to cause serious
explosions.
In the absence of sodium, lithium, and magnesium salts, the
electrolytic extraction of the metal is effected under almost
identical conditions with those observed for sodium, but the bath
employed consists of a mixture of the haloid salt of potassium
with that of an alkaline-earth metal. It will suffice, therefore,
to refer to a few methods especially proposed for the extraction
of potassium.
Matthiessen's observation* that potassium alone resulted from
the electrolysis of a mixture of calcium and potassium chlorides
* Liebfig's Ann,, 1865, vol. xciii., p. 277.
Digitized by VjOOQ IC
72 ELECTTBIC SMELTING AND REFINING.
was interesting ; but the advantages obtained from the redaction
in the fusing point of potassium chloride by the addition of the
calcium salt are more than balanced by the difficulties which
always accompany the electrolysis of mixtures containing calcium
chloride.
Linnemann's method* depends upon the decomposition of
cyanides. The current passes from a carbon plate to a pointed
carbon rod immersed in potassium cyanide, which is kept melted
in a crucible. If the temperature be so controlled that the
upper surface of the bath remains solid, the separated potassium
will accumulate under the crust. Ilie high cost of the raw
material, however, renders unnecessary any remarks upon the
practical utility of the process.
All other processes are already described under "sodium," and
reference may be made to them in that chapter.
Uses of Potassium. — On account of the considerable violence
of its reactions, the hitherto dangerous method of production,
and the high price of potassium salts, the applications of potassium
have remained so insignificant that there is practically nothing
to be written concerning its use industrially.
* Jourfu fiir pi-akt, Chem., 1848, vol. bcxiii., p. 416.
Digitized by VjOOQ IC
CAlXnUH, STRONTIUM, BARIUM. 73
THIRD GROUP.
CHAPTER V.
CALCIUM, STRONTIUM, BARIUM.
Ooourrence of Caloium, Barium, and Strontium in Kature.
— ^Tbe chemical properties of the metals of this group are sach
that only their salts are found in nature. Among haloid salts
the principal is flaor spar, CaF2 3 all three metals oommonlj
occur as sulphates : calcium, in gypsum, OaSO^ + 2H2O, and
anhydrite, CaSO^ ; strontium, in celestite, SrSO^ ; and barium,
in heavy spar, BaiSO^ ; they are often met with as carbonates :
-in calc spar, marble, chalk, and limestone, CaCOg; in strontianite,
SrGOg ; and in witherite, BaC03 ; and, finally, they may all be
found as phosphates, borates, and silicates, but in these calcium
occurs the most abundantly.
Calcium (Oa'^; atomic weight = 40; specific gravity =1*85). —
The metal, after fusion and when absolutely free from nitrogen,
is lustrous, silvery white, and brittle ; it may be 'cut with a
knife, is less malleable than the alkali metals, and shows a
crystalline fracture. It will scratch lead, but not calc spar. It
crystallises in hexagonal plates, or rhombohedra (mostly very
regular), and sometimes in dendritic or stellar formations. The
melting point of the metal when heated in a lime vessel in
vacuo is 760*" 0. With hydrogen at a red heat, calcium forms a
white crystalline hydride, CaHg. With the halogens it combines
only when heated. It burns when gently warmed in air, and,
when heated to 300^ 0. in a current of pure oxygen, it oxidises
with so great an evolution of heat that the lime produced by the
reaction melts and, in part, volatilises. The formation of a
peroxide has not been observed. In a current of nitrogen at a
red heat it combines to form the yellow nitride Ca3N2' ^^®
metal, therefore, when heated, can fix the two principal con-
stituents of the atmosphere. When heated, it combines directly
with sulphur, selenium, tellurium, as also with phosphorus,
arsenic, bismuth, carbon and silicon; whilst the sulphide is
only decomposed by acids, evolving hydrogen sulphide in the
reaction, the compounds with the second group of elements just
enumerated are all decomposed by water. Thus the phosphide,
when placed in water, yields spontaneously-in flammable hydro-
Digitized by V^jOOQ IC
74 ELECTRIC SMELTING AND REFINING.
gen phosphide ; the arsenide, gaseous and black solid hydrogen
arsenides; the antimonide, only solid black antimonetted hydro-
gen; and the bismuth compound, hydrogen and a black
insoluble compound; whilst the silicide, only attacked slowly
by water, rapidly evolves with dilute acids a mixture of
hydrogen and hydrogen silicide. It is noteworthy that the
direct union of calcium with carbon to form carbide commences
at a dull red heat, but that the reaction temperature rises so
high that the resulting carbide fuses. Calcium dissolves in
melted sodium (but not in potassium) and crystallises out again
on cooling. It alloys with magnesium, zinc, nickel and tin.
The crystalline amalgam formed by triturating calcium with
mercury in an atmosphere of carbon dioxide at once takes up
oxygen and nitrogen when exposed to the air, and becomes
covered with a grey-black film; it is indifferent to hydrogen
at temperatures below 400° 0., but, when heated in an atmos-
phere of nitrogen, calcium nitride is formed. Calcium, at a
red heat, readily deposits the alkali metals from their chlorides
or fluorides, but not from their iodides. It reduces the calcium
chloride and iodide to sub- salts. Further, calcium reacts, when
heated, with sulphur dioxide, nitric oxide, phosphorus pent-
oxide, boric acid, silicic acid, carbon dioxide, carbon monoxide,
hydrogen sulphide, boron trichloride, acetylene, ethylene,
methane, and the compounds of hydrogen with the halogens.
It reduces fuming and anhydrous sulphuric acid in the cold,
and is dissolved by hydrated acids. It is indifferent to am-
monia in the cold, but, when heated in it, it decomposes the
gas, forming a mixture of nitride and hydride. It forms a
solid white compound (which has not been further examined),
when cooled to 40"* C, with aqueous ammonia.
Strontium is also a soft white metal, the other properties of
which will be studied presently.
Barium. — ^Keeping in mind the later determinations of the
properties of calcium and strontium, the older statements as to
the properties of barium (a metal which has not been produced
pure, as yet, in such great quantities as have calcium and
strontium) must be accepted with caution.
Extraction of the Metals. — The separation of the three
metals of the alkaline earths from their oxides or haloid salts is
attended with very great difficulty as compared with that of the
alkali metals. Davy * was the first to accomplish the electro-
lytic decomposition of the alkaline earths. He moulded the
moistened hydroxide, either alone or mixed with mercuric
oxide, into the shape of a small cup, rested this on platinum
foil, which served as the positive pole ; he then poured mercury
into the cup of hydroxide, and made this the negative pole.
An amalgam was thus produced, from which the mercury could
* PhU, Trans., London, 1808, p. 335.
Digitized by LjOOQ IC
• CALCIUM, STRONTIUM, BABIUM. 75
be expelled by heat. According to subsequent investigations,
it would seem that Davy did not produce a perfectly pure
metal, free from mercury and silicon. Bunsen and Mattbiessen,
by the decomposition ot the chlorides, were the first to prepare
the alkaline-earth metals in the pure condition. For a long
time fruitless attempts were made in Bunsen's laboratory to
electrolyse the chlorides of these metals in the apparatus which
had given such good results in the extraction of magnesium;
and these experiments were continued until, in the course of
his researches on metallic chromium in 1854, Bunsen discovered
the cause of his previous failures. He at that time published
the following noteworthy observation,* which has since con-
ditioned the success of a large number of electrolytic decom-
positions : —
"The density of the current used for electrolysis — that is, the
relation of current volume to electrode area — exerts a most
important influence on its chemical effects. The power of the
current to overcome (chemical) affinities increases with this
density. . . . Of no less importance is the relative mass of
the constituents of the electrolyte through which the current
passes."
In the same treatise, Bunsen showed that by using a sufficient
current-density it is even possible to separate calcium and
barium from boiling concentrated solutions of their chlorides,
acidified with hydrochloric acid. As negative pole, he used an
amalgamated platinum wire immersed in the mass under treat-
ment. The latter was placed in a clay cell, around which stood
a carbon crucible, partly filled with hydrochloric acid, and set
within a porcelain crucible, the whole arrangement being heated
in a water bath. The carbon crucible was used both as anode
and as electrolyte cell. Great difficulty was experienced in
the electrolysis of calcium chloride by this method, because,
even after a few minutes, the electrode became covered with
a crust of lime which interrupted the current. It was, therefore,
necessary to lift out the platinum wire, remove the rapidly-
dried coating of amalgam, and then re-amalgamate it. Barium
amalgam may readily be obtained in quantities of about 15
grains by using crystallised barium chloride, made into a paste
with weak hydrochloric acid, at a temperature of 100° 0. For
these exi>eriments a current-density of about 1 ampere per sq.
mm. [645 amperes per sq. in.] of cathode surface is necessary.
Matthiessen's Experiments. — ^As a result of these experi-
ments, Mattbiessen,! working in Bunsen's laboratory in the
year 1855, obtained the metals of the alkaline earths in the
pure condition by the direct treatment of the fused chlorides.
He writes that : " Although the reduction is easy, it is difficult
♦ Pogg. Ann,, 1854, vol. xci., p. 619.
t Liebig'a Ann., 1865, vol. xchi., p. 277.
Digitized by LjOOQ IC
76 ELECTRIC SMELTING AND REFINING.
to obtain the reduced metal in coherent masses, and to separate
it from the fused substance. The deposited metals for the most
part rise to the surface, by virtue of their low specific gravity,
before they have grown to globules of appreciable size; and there
they burn so quickly that it is almost ini possible to collect them.
If the end of the electrode be provided with a bell-shaped attach-
ment of glass or burnt clay, in order to collect the metal, the latter
reduces a small quantity of silicon, which separates in the form
of a black powder, and prevents the metal from uniting into a
regulus."
He proposed three methods of overcoming this difficulty.
The^r*^ method, by which, however, only a very impure metal,
or rather an alloy, is obtained, consists in the use of a platinum
wire as the negative pole. The alkaline-earth metal thus
becomes alloyed with platinum, and so acquires a specific
gravity sufficiently high to enable it to sink through the fused
chloride. After cooling and breaking up the mass the metal is
obtained in the form of large grains.
A second plan is to melt together two chlorides in simple
molecular proportions, by which means a double chloride is
produced of so fusible a nature that easily vaporised metals,
like potassium or sodium, may be separated in it without
volatilising. A porcelain crucible is used for the experiment,
and if the temperature be so r^ulated that a solidified film
forms on the upper surface of the mass around the negative pole,
this film, on cooling, will be found laden with metallic grains,
which may be readily separated from the surrounding material
by crushing the whole of the cooled contents of the crucible
under petroleum in a mortar. The metal will be visible in the
shape of small plates or laminse amid the ])ulverised mass.
The third way is based on the separation of the metal
immediately beneath the surface of the melted chloride by an
electrode consisting of a pointed iron wire, which serves to
remove the metal floating on the upper surface as well as that
clinging by adhesion to the point of the iron wire. By this
method the liberated metal is sufficiently protected from oxida-
tion by a thin varnish-like film of molten chloride to enable the
particles to increase to the size of mustard seed.
Eztraotion of Caloium. — The following method of reduction
is exceedingly uncertain in operation ; but when successful it is
capable of yielding fragments of calcium somewhat larger than a
pea. '*A mixture of two molecular weights of calcium chloride
with one equivalent of strontium chloride, and with sal-
ammoniac, is melted in a Hessian crucible until the last-named
constituent has volatilised ; an iron cylinder, to be used as the
positive pole, is then placed in the melted salt mixture, and
within this is immersed a narrow clay cell, about 4 inches
long, previously heated to a red heat. The clay cylinder is filled
Digitized by V^OO^ It!
CALCIUM, 8TB0NTIUM, BARIUM. 77
with the same fused mixture of salt, and serves for the recep-
tion of the negative pole, which may be either an iron wire of
about the thickness of a knitting-needle or a carbon rod. If it
be arranged that the level of the fused chlorides in the clay
cell is about ^ to 1 in. higher than that of the liquid in the
crucible, the heating may, with ease, be so regulated that a
solid crust forms only on the surface of the liquid in the cell ;
and the metal will then collect beneath the crust without coming
at all in contact with the walls of the cell. With the current
from six zinc-carbon elements, such as JMatthiessen used in nearly
all his experiments, a large quantity of reduced calcium may be
obtained after the space of from half an hour to an hour. Only
once, howevpr, were a couple of pieces, which had been fused to
the size of a pea, obtained by this process ; and almost always the
metal was found to be distributed in the form of a fine powder
in separate parts of the cooled and crushed slag.
'^ Calcium may be obtained in small molten globules more
simply and with greater certainty by fusing the mixture in a
smsdl porcelain crucible heated, as in igniting precipitates, over
a spirit-lamp or between lumps of charcoal, and then passing the
current through the electrolyte from a carbon electrode, which
should be as large as possible, to a fragment of pianoforte wire
only two lines long (No. 6 size), which is connected to the
negative pole by a stouter wire reaching as far as the upper
surface of the liquid. A crust of solidified salt should be
allowed to form upon the surface around this wire, which may
be removed from time to time ^at intervals of three minutes)
in order that the crust and metal may be detached in a mortar.
*<The metal may be reduced in a similar way by so stirring the
upper surface of the melted chlorides with the point of the wire
for from one to two minutes, that a glow, produced by the current
itself, is observed around the point. Or the point of the wire
may, at intervals, be first dipped into the liquid and then raised
until a small electric arc is observed at the surface ; this will
cause alternations of cooling and strong heating, which should
effect the fusion of the metal that has been deposited in
pulverulent form.
"Calcium is not reduced from its chlorine compound by
sodium or potassium. By fusing together one molecular weight
of calcium chloride and two of sodium chloride, or equal
molecular weights of calcium and potassium chlorides and
sal-ammoniac, a double chloride is obtained, which melts at
a temperature below the vaporising point of either sodium
or potassium. If such a [sodium] mixture be electrolysed at
a bright red heat with a carbon anode and a pointed iron-
wire cathode, a number of comparatively large globules of
sodium will be seen to form at the end of the wire (which
should only just dip into the liquid) and to rise to the surface^
Digitized by LjOOQ IC
78 ELECTRIC SMELTING AND REFINING.
where they will slowly circulate and burn. If such a globule
can be removed from the bath, it will be found to contain
scarcely a trace of calcium."
Matthiessen wds of opinion that if it could be arranged to
heat the under part of the crucible containing the mixture
2CaCl3 + SrClg to a very high temperature, without fusing the
crust on the surface, the calcium might be melted to a regulua
below, and this process would then be preferable to any other.
These conditions of irregular heating are easily fulfilled by using
a crucible of the form recommended for the extraction of mag-
nesium (p. 9), but the fusion. of the calcium to a regulus has
not yet been accomplished in this way.
Extraotion of Strontium. — According to Matthiessen's
accounts^ the production of strontium appears to be less
difficult. A small crucible and a clay cell placed within it are
filled with anhydrous strontium chloride mixed with a little sal-
ammoniac, so that the surface of the fused mass in the cell may
stand at a higher level than that in the crucible. The clay
cell is surrounded by a cylinder of iron which serves as the
positive pole, and into the cell is dipped a short and very thin
piece of iron wire ; the latter is ^eustened to a thicker wire
which is enclosed in a pipe-clay tube, reaching as far as the
projecting piece of thin wire beneath. If the temperature be so
regulated that the mass in the cell is covered with a crust of
solidified material, the strontium will collect under the crust
(and without touching the side walls) in fragments that may
weigh as much as 7 or 8 grains. Barium, however, can be
obtained only as a fine powder by a process analogous to this.
The statement in Gmelin-Kraut's Hcmdkuch der anarganischen
Chemie f that Matthiessen had obtained metallic barium in
globules the size of mustard seed, adhering to an iron electrode
by the electrolysis of fused barium chloride, is founded on an
error. The sources of information | quoted contain no such
statement.
So far Bunsen's and Matthiessen's work. Hiller recommends
the apparatus described in the chapter on lithium (p. 23) as
specially suitable for strontium reduction. Nevertheless, satis-
factory as it has shown itself to be for lithium, it is useless for
the production of strontium in pieces of appreciable size. In
most cases it will be found, on breaking the crucible, that by far
the greatest part of the metal separated at the negative pole has
sunk to the bottom, and that it has there reduced the walls of
the crucible to silicon and aluminium, or even that it has
* Liebig-Kopp'a JahreshericJU, 1866, p. 323 ; and Quart, Joum, Ckem,
.Society, 1866, vol. viii., p. 107.
t Handbuch der anorganiichen Chemie, vol. ii., p. 256 (Ed. 1886).
XLMng's Ann,, 1866, voL zciii., p. 277; and Liebig-Kopp's Jahrea*
bericht, 1866, p. 320.
Digitized by VjOOQ IC
CALCIUM, STBOKTIUM, BARIUM. 79
become re-oxidised at the positive electrode owing to its having
been carried forward by the circulation of the liquid caused by
the necessarily high current-density used.
The method originated by A. Feldmann* must be regarded as
retrograde. He proposes either to mix the single or double
haloid salts of the alkali- or alkaline-earth metals with the oxide
of a more electro-positive metal, or else to melt the oxide of the
metal that it is desired to obtain with the haloid salts of one (or
of several) more electro-positive metals, and then to decompose
the fused mixture by means of the current. The metal to be
deposited should be present in the electrolyte either as oxide
alone, or as haloid salt alone, but not simultaneously in both
forms. A sufficient explanation has been given of the disadvan-
tages resulting from the presence of oxides in applying the
methods which have hitherto been described for the electrolysis
of the fused haloid salts of alkali- or alkaline-earth metals ; and
there is consequently no need to re-state the case as against
Feldmann's process.
Eztraotion on a Commeroial Soale. — Up to a date a few
months before the publication of the German original of this
book, the view was indisputable that there was apparently no
more difficult task presented in the elecfcro-metallurgical industry
than the manufacture in large quantities of the met<als calcium,
strontinm, and barium. Yet, on the other hand, almost all the
conditions that must be observed in the solution of this problem
may be learned from the writings of Bunsen and Matthiessen,
by whom they were clearly enunciated. There are necessary : —
1. A high current-denBlty (500,000 to 1,000,000 amperes per sq. metre
[323 to 645 amp. {ler sq. in.]).
2. A high temperature at the cathode.
3. A low temperature for the electrolyte*
The first and second requirements may very easily be brought
into agreement, but not so the second and third. Yet it is
necessary that they should be fulfilled, for the metal under
deposition must be united as quickly as possible into cohesive
masses ; finely-divided metal is very liable to form sub-chlorides,
and so, passing into the melt again, to be burned at the anode ;
a high temperature and a rapid separation of large quantities
of metal at the cathode are indispensable, and, as already
explained, both these conditions may be fulfilled by applying
a high current-density at the cathode. But at the same time—
and herein lies the secret of the technical difficulty — the
electrolyte in the immediate vicinity of the cathode must be
kept at the lowest possible temperature, or else the haloid
compound will redissolve even large globules of metal, and so
convey the metal to the anode [as the hypothetical sub-chloride].
The destructive action of fused and highly-heated alkaline-
* German Patent 50,370l [English Patent 9,783, July 5, 1888.]
Digitized by LjOOQ IC
80
ELECTRIC SMELTING AND REFINING.
earth salts upon all materials that may be employed in the
construction of containing vessels needs no comment. Without
having to read much between the lines, it is possible to find all
the necessary suggestions upon these points in the writings of
Bunsen and Matthiessen, to which reference has been made so
frequently. The very thin iron cathode wire, dipping only just
beneath the surface of the electrolyte, must of necessity be
strongly heated by any current of high density — and in this
way the second of the above requirements is fulfilled. On the
other hand, it has frequently
been insisted upon that the bulk
of the electrolyte should be
maintained at a temperature so
low that, if possible, a solid crust
may be formed on the upper sur-
fisM^e. The form of the experi-
mental apparatus described in
the original papers, as well as
the whole method of applying
it, completely excludes the possi-
bility of using such a plant for a
profitable iostaUation on the
large scale, as, indeed, the
authors themselves admitted.
At the moment of separation
the metal has a very high tem-
perature, and its specific gravity
being for the time lower than
that of the surrounding salt, it
rises rapidly to the surface and
most of it is burned. This diffi-
culty may be met by using an
apparatus such as that first
described by Hiller (p. 23). But,
unfortunately, the relation be-
tween the volumes and weighta
of the metal and the fused salts
is easily altered. As soon as any
considerable quantity of metal
has separated on the thin wire electrodes and has, in conse-
quence, led to an increase in cathode area, the temperature at
this point must necessarily fall. The specific gravity of the
metal is thus increased by so much that, before they have time
to solidify, drops become detached from the electrodes, and sink
through the fused salt ; and that which is not dissolved while
falling finds an easily reducible silicate (porcelain) when it
reaches the bottom. To counteract this evil Borchers, as men-
tioned in the first German edition of this book, gave the cathode
Fig. 67. — Borchers' experimental
apparatus for alkaline-earth
metal reduction (^-scale).
Digitized by
Goog
It:
CALCIUM, STRONTIUM, BARIUM. 81
the form ehown in Fig. 57. Within a wide iron tube having
a ba]ged-in bottom, R, is placed a narrower tube, r, reaching
nearly to the bottom. This narrow tube is used to introduce
into R a stream of cold water which, becoming warmed,' maj
find an escape through the tube, S. At the bottom of the
tube, R, and on the outside, is attached an iron pin-like pro-
jection, which must be of such a size that it may become
heated to a bright red heat by the electric current passing
through it on immersion in the electrolyte; it is necessary
that it should thus be heated to a temperature higher than
that of the surrounding salt. The tube, R, is sunk to the depth
of only 4 to ^ of an inch in the fused mass ; and care must be
taken that, both before the immersion and during the whole
operation, a current of water is circulating through r, so that the
bottom of the tube may become coated with an insulating crust
of solidified salt, and that the whole of the current may thus be
concentrated upon the surface of the projecting pin. The metal,
as it drops from the electrode, falls into the iron crucible, T,
which is placed beneath R, but is insulated from it. The copper
wire, K, soldered to the upper end of the tube, R, makes the
necessary electrical connection with the generator. The bottom
of the tube, R, is made concave exteriorly, in order to prevent
the escape and subsequent combustion (already alluded to) of
the very hot, and therefore relatively light, metal at the moment
of production. The incrustation induced by water cooling, and
the reaulting insulation of that portion of the tube, R, which is
immersed in the solution, have for their object the reduction of
the conducting area of the cathode. As soon as the reduced
metal falls from the surface of the pin-shaped cathode, it must
be as quickly as possible withdrawn from the active circulation
of liquid between the electrodes that results from the high
current-density employed. This is accomplished by the use of
the crucible, T, which is suspended at a short distance below the
negative electrode. The metal there sinks through the layer of
tranquil liquid, protected by the walls of the crucible, and
rapidly solidifies at the bottom.
The yields of metal were very small, and even qualitatively
but little satisfactory.
Later experiments have shown that the greater part of the
metal already separated rapidly passed into solution again,
whether it had first sunk into the crucible, or risen to the
surface, or had remained caught in the wire gauze. This action
is doubtless favoured by the fact that the wire and the crucible
form intermediate electrodes, and so act as anodes on the side
upon which the metal from the cathode has fallen or to which it
is clinging.
Moisaan* has since found that the electrolysis of calcium
*Comptu RenduBy 1808, vol. cxxvi., p. 1758.
Digitized by LjOOQ IC
82
BLFCTRIC SMELTING AND RBFINING.
iodide offers fewer difficulties than the methods previously
described. The electrolysis is effected at a dull red heat, using
a nickel cathode and a graphite anode, surrounded by a porous
FigB. 58 and 59. — Borchers and
Stookem*8 oaloioin extraction
furnace.
Fig. 59.
vessel. The fused salt conducts the current well. The tem-
perature is maintained at the melting point of calcium iodide by
the agency of the current^ and iodine vapour is rapidly given off.
Digitized by V^jOO^ It^
CALCIUM, STRONTIUM, BARIUM.
83
The metal is obtained in the form of small white crystals or
globules. No detailed description of the apparatus is, however,
available.
New experiments have lately been made in Borchers' labora-
tories at Aachen, in which he, with his pupil, L. Stockem, has
succeeded in overcoming entirely the difficulties in the produc-
tion of calcium, and, for the most part, those met with in the
extraction of strontium.
Borohers and Stookem'a Caloinm ProoeBs. — Owing to the
success of this process, the production of calcium is now one of
the simplest of electro-metallurgical processes, and the produc-
tion of this metal in any desired quantity is possible at a very
low price.
The apparatus first used for the puipose, as shown in Figs.
58 and 59, has not proved itself very suitable practically for the
extraction of calcium. It is here illustrated, however, because
it is an early type of an apparatus for the extraction of
Fig. 60. Fig. 61.
Borchers and Stockem's modified calcium extraction furnace H n.8.).
strontium. The drawings show at a glance that the principles
of construction used in Borchers' latest form of apparatus for
the extraction of alkali metals have been retained Only it will
be seen in Fig. 59 that a somewhat smaller cathode is used, and
that a space is provided for the collection of the metal that
sinks to the bottom. The possibility of separating calcium in
considerable quantities from electrically-heated chloride was
established by means of this apparatus ; the condition of the
metal separated, however, showed that the form of the apparatus
was not entirely satisfactory. For example, at a moderate red
heat^ the calcium separated out at the bottom in a bulky form,
so that a bridge was rapidly formed between the anode and the
cathode at the bottom. On raising the temperature, the sepa-
rated metal quickly re-dissolved in the fused bath in the form of
Digitized by V^jOOQIC
84 ELECTRIC SMELTING AND REFINING.
a sub-oarbide of calcium, described by Stockem. The largest
amount of calcium was consequently obtained, as was proved by
later experiments, when electrically-fused calcium chloride was
electrolysed at a moderate red heat, under which conditions the
metal sponge could be removed from the bath as it separated at
the cathode. In order to facilitate this, the cathode was raised
somewhat and the apparatus was worked with an open top under
a good' draught, no attempt being made to recover the chlorine.
The apparatus, therefore, was in the form illustrated by Figs*
60 and 61. Like the sodium apparatus, it consisted of a
cylindrical melting vessel, composed of carbon blocks arranged
like the staves of a cask, and serving at the same time as
anode. A bottom of fluor spar, which is prevented from
melting by a cooling chamber beneath, is stamped into this
cylinder. The cathode consists of an iron rod, and, without any
intention of so cooling the rod, it was found most convenient to>
connect it with the current mains by screwing it into the
cooling vessel. It was made so thin that any cooling effect from
below was more than compensated for by the heating action of
the current. With an apparatus of the dimensions shown in
Figs. 60 and 61, a current of from 120 to 140 amperes at an
average pressure of 12 volts was used. A clay cylinder was
placed between the support of the cathode and the carbon
cylinder for the purpose of insulation.
The process was conducted as follows : — ^A layer of .fluor spar
was first stamped into the space within the carbon cylinder upon
the bottom formed by the clay cylinder and the cooling cham-
ber. Calcium chloride in the form of powder was then placed
on the fluor spar until the cathode was completely covered. Twa
or three carbon rods were then embedded [radially] in the top
layer of the chloride, as -shown in Fig. 61, in which the furnace
is viewed from above. On switching on the current the calcium
chloride melted, first in the immediate neighbourhood of the
carbon rods, and afterwards throughout the whole of its upper-
most layer. The carbon rods were then immediately, and as far
as possible simultaneously, removed from the fused mass, which
now conducted the current and was at once subject to electro-
lysis. The calcium separated as a soft spongy mass in the
neighbourhood of the cathode. Chlorine was, of course, evolved
at the anode, and, in these experiments, was allowed to escape.
At first the spongy metal was removed from the mass with the
aid of an 'iron spatula or ladle, and was at once plunged into
petroleum. Much chloride was necessarily left in this sponge^
which, however, when saturated with chloride, contained nearly
60 per cent, of free calcium and showed a silvery- white metallic
lustre on freshly-cut surfaces. In later experiments the sponge
was consolidated with the aid of broad-ended tongs fashioned
somewhat in the shape of a gauffering-iron, before removing it
Digitized by VorOO^ It!
CALCIUM, STRONTIUM, BARIUM,
85
from the bath. In this way the entangled calcium chloride was
removed until but a small proportion was left ; and, at the same
time, the calcium sponge was so welded together that the mass,
after removal from the bath and cooling, presented a quite
homogeneous surface, showing a silvery-white lustre after it had
been cut through. The crude metal contained about 90 per
cent of metallic calcium, and could be readily melted to pure
calcium in closed iron or nickel vessels. Stockem found that,
after thus remelting the calcium, the layer of salt left on the
surface of the metal, after the cooling of the melting vessel, was
not calcium chloride but a sub-chloride, probably CaOl. This
.salt formed transparent red crystals belonging to the monoclinic
or triclinio systems. The crystals could not be accurately
.measured, partly by reason of their small size and partly because
they were so readily decomposed that, even while under exam-
ination by the microscope, they were attacked by the moisture
of the air, with evolution of hydrogen. But their crystalline
form and colour support the view that the substance is not
merely a solution of calcium in calcium chloride, but is an
actual sub-chloride. The formula could not, however, be ascer-
tained, owing to the presence of calcium and calcium chloride
between the crystals of the sub-chloride.
Production of Strontium. — The results obtained in the
electrolysis of calcium chloride naturally led to the application of
the method to the treats
ment of the chloride
of the other alkali-earth
metals. Strontium
chloride required the
use of a currentdensity
nearly twice as great
as that used for calcium
chloride, in order that
it should remain fluid
and that the electro-
lytic separation of the
metal might be effected.
The temperature re-
quired was also higher,
and, indeed, could not
be maintained so low
that the metal should be deposited in the form of a sponge
which could be lifted from the cathode; but, at a red heat
somewhat brighter than that used in the electrolysis of calcium
chloride, it separated in the form of globules which at times
floated to the surface and then quickly sank again, since
their specific gravity was probably but little different from
that of the bath. The apparatus must therefore be so arranged
Digitized by VjOO^ It^
Fig. 62. — Borchers and Stockem's strontium
extraction furnace.
86 B^HCTRIO SMELTING AND REFINING.
that the metal collects on the hottom and is, as fiur as pos-
sible, cooled there, in order to minimise its tendency to re-dis-
solve in the chloride, a tendency which is as great as that of
calcium to dissolve in the calcium chloride. This was ultimately
effected in the apparatus shown in Fig. 62. The cylindricid
anode was retained as the side-wall of the melting vessel ; the
cooling compartment at the bottom was, however, widened, and
was of course covered with a layer of a strontium salt, the
chloride being found to be the most suitable salt for the purpose.
A thin iron rod screwed into a thicker rod passing through the
middle of the cooling jacket was used as cathode, and projected
upwards only to the lowermost part of the carbon cylinder.
Under these conditions, the metal separating at the cathode was
but little affected by the chlorine evolved at the anode, and,
sinking to the bottom, rested on the crucible of cooled strontium
chloride. In this way, after stopping electrolysis and allowing
the charge of salt to solidify, globules of strontium were
obtained of all sizes up to about 10 mm. The metal was readily
separated from the slag on account of its softness; for, on
crushing the charge, the globules of strontium were flattened and
could be separated from the crushed mass by means of a sieve.
In contradistinction to the statements made in the earlier
literature of the subject, it was found that strontium, like
calcium, is a silvery-white and very soft metaL
Froduotion of Barium. — Stockem has made many attempts
to obtain barium in the apparatus which had been successfully
used for calcium and strontium, but so fiu: without result. The
barium is separated in so fine a state of division that it, for the
most part, floats to the anode in suspension in the fluid charge
(which is in constant motion) and so becomes burnt. It is true
that the solidified charge, especially from the lower layers,
evolves hydrogen when thrown into water, probably because
the salt is permeated with minute particles of metal ; but the
action is much more sluggish than in the case of calcium or
strontium salts taken from the neighbourhood of the cathode
after electrolysis for the separation of these metals.
Beduotion of the Oxide. — The oxides of the alkali-earth
metals, like all the other oxides, as stated in the first [German]
edition of this work in 1891, are capable of reduction by means
of electrically-heated carbon ; they yield no metal, however, but
only carbide in this way. £ven if a mixture be used containing
no more carbon than is required by the equation
CaO + C =r Ca + CO,
a quantity of carbide is produced corresponding to the supply of
carbon, and this carbide mixes with the residual unchanged
oxide, apparently in all proportions, and without any reaction
leading to the separation of the metal.
Digitized by VjOOQ IC
BBBTLLIUM. 87
If the carbide be subjected to a very high temperature for a
long time it is decomposed and carbon is left behind as graphite^
and the calcium, being volatile at this temperature, passes off;
but up to the present time no apparatus has been found which
has led to the metal being so obtained. In experiments of this
kind it is often found that the residue contains finely-divided
caldam, which is recognisable by the rapid production of calcium
oxide or hydroxide as soon as air gains access to it.
The carbide will be treated of at the end of the book, with the
carbides of other metals.
TJseB of the Metals. — It is obvious that the great difficulties
which have, until this year, stood in the way of the separation
of the alkali-earth metals in quantity, have prevented these
metals irom finding any practical use. There is now, however,
a good prospect that they — or, at least, that calcium — may find
many applications in scientific and technical work. Calcium is
chemically so active that its assistance will be welcome in many
branches of the chemical industry. In metallurgy it will prove
to be a very useful reagent for reductions and precipitations, as
well as for the elimination of sulphur and phosphorus, and
possibly also of carbon from crude metals ; and it may also be
used for thermo-metallurgical reactions. It will, however, be
especially useful in organic chemistry ; for in its chemical pro-
perties it is very similu* to the alkali metals, but in its reactions
it is much more violent, and the calcium passing into the
solutions is very easily separated out again.
CHAFIER VI.
BERYLLIUM.
Ooourrenoe in Nature. — ^The metal occurs almost exclusively
in combination in silicates, of which the principal is beryl, a
double silicate of aluminium and beryllium wiUi the formula
(BeO)3 . AI2O3 . GSiOg. The emerald is a form of beryl that is
especially prized as a precious stone.
Properties of the Metal. — Beryllium (Be*; atomic weight =
9; specific gravity *= 1*64) is a lustrous, white, soft, malleable,
and tough metal, with a melting point approximating 800* O.
In the form of powder, thin sheet) ribbon, or wire, 5ie metal
bums in the air like magnesium, the halogens, sulphur, drc.
Beryllium withstands the action of water better than magnesium,
but it is equally readily attacked by acids; in salt solutions
it slowly oxidises with evolution of hydrogen.
Digitized by LjOOQ IC
88 KLECTRIC BMELTINO AND REFINING.
Extraction of the Metal. — Numerous methods are given
for the treatment of beryl. The method of fluxing varies but
little from that of other silicates, being efifected by means of
alkaline carbonates. The fusion is afterwards treated with
sulphuric acid to render the silica insoluble; and there then
results a solution of beryllium, aluminium, and iron salts. The
bulk of the aluminium is then separated, either as alum by
the addition of potassium sulphate, or as aluminium fluoride by
boiling with cryolite or other fluorides. If the quantity of iron
present be neither too small nor too great, it is carefully pre-
cipitated by the addition of soda, with the precaution that any
ferrous salts shall be previously peroxidised by the addition of
alkaline chromate, chloride of lime, or other suitable oxidising
agent. On the further addition of barium carbonate, or an
alkaline carbonate, a precipitate of aluminium hydroxide will
be formed which will carry some beryllium down with it ; but
the latter may afterwards be separated as beryllium hydroxide
by the addition of ammonium carbonate solution to the pre-
cipitate. The remaining solution should contain only beryllium
sulphate.
Eleotrolysis of Beryllium Compounds. — ^The above solu-
tion is as little suitable for use as any other aqueous solution of
a beryllium salt, and it can be readily understood that, of the
Compounds that can be obtained anhydrous, only the haloid salts
are utilisable. When, therefore, the metal has been obtained in
solution as a sulphate by the above method, it must be converted
into chloride, or into some other suitable haloid salt. With this
object in view, the beryllium is either precipitated as carbonate
by the addition of an alkali-metal carbonate, and the carbonate
is then re-dissolved in a halogen acid, or the sulphate is decom-
posed by means of barium chloride or some other convenient
haloid salt of barium. The solution of beryllium chloride
obtained by one of these methods is then evaporated to dry-
ness, with the addition of a chloride of an alkali- or alkaline-
earth metal* (excepting magnesium chloride or calcium chloride),
or of some other haloid salt of an alkali metal, most conveniently
the fluoride ; ammonium chloride or other haloid compound of
ammonium is also added to prevent the formation of oxides.
In this way there is obtained a mixture of salts that is fusible,
and is also a good conductor of electricity; and from this the
beryllium may be separated electrolytically, in the same way
as magnesium or lithium is reduced, although, according to
Hampe's experiments, beryllium chloride is reported not to be
an electrolytic conductor. There is practically nothing to be
said concerning the electrolysis itself; both the apparatus and
the manner of conducting the process are the same as in the
case of magnesium. Care should, however, be taken that the
temperature never exceeds that at which the metal fuses, other-
Digitized by V^OOy It!
BERYLLIUM. 89
wise the beryllium will tend to alloy with iron. On account of
the higher temperature that is used, as compared with that in
the electrolysis of camallite, it is also of importance that the
melting vessel be made of good wrought iron.
Iiebeau's Prooess. — ^The efficacy of this pixxsess, which the
author had described in the second German edition of this work,^
was confirmed by the experiments of Lebeau,t who obtained the
metal from the double fluoride of beryllium and an alkali metal.
According to his statements, the double salt was melted in a
nickel crucible, which also served as cathode, and was then
electrolysed, using a rod of graphite carbon as anode, without
the application of external heat, the current employed being
one of from 6 to 7 amperes and from 85 to 40 volts. The metal
precipitated on the cathode in the form of crystals which could
readily be detached from it quite pure. Beryllium alloys, for
example the beryllium bronze which had been described some
time before, could be made by employing a carbon crucible with
another metal as anode.
Lebeau I ascertained, also, from his experiments that the sub-
stance obtained by Berzelius by dissolving beryllium hydroxide
in hydrofluoric acid, evaporating dry, and .heating the residue in
air, is not beryllium fluoride, as originally stated, but a beryllium
oxyfluoride, which is stable at a red heat and has the formula
dBeFj . 2BeO. The anhydrous fluoride (BeFg) cannot, therefore,
be prepared in this way, but must be obtained either by heating
the above-described oxyfluoride in a current of hydrogen
fluoride, or by heating the double ammonium-beryllium fluoride
(BeF^ . 2NH-F), first described by Marignac, in a current of
carbon dioxide ; this is identical with the fluoride obtained by
heating beryllium carbide in fluorine or in hydrofluoric acid gas.
It is a transparent, glassy substance, with a specific gravity of
2-1 at 15* C. ; it softens on heating, and flows freely at about
800* C, subliming in small crystals at this temperature. It is
very hygroscopic, dissolves easily in water and in hot alcohol,
and is converted into the oxyfluoride by heating in oxygen, or
into the sulphate by treatment with sulphuric acid. It is
decomposed, with separation of beryllium, by the alkali metals
and by magnesium, but not by aluminium.
Iiiebmann'8 Process. — Liebmann§ has proposed the direct
electrolytic treatment of native beryllium silicate with the
addition of fluorides, in order that the silicon of the silicate
might combine during electrolysis with a portion of the fluorine
from the fluoride, and so be(X)me volatilised ; the beryllium was
to be obtained in the pure metallic state, or, better, in the form
* EUietrO' MetcUlurgie, p. 30, Leipzig, 1894-5; cf. Electric Smelting and
Bejimng, let ed., 1897, p. 41.
t Gomptes Bendiut, ».8»8, vol. cxxvi., p. 744. * Ihid., p. 418.
g German Patent, Nos, 101,326 and U)4,632. '
Digitized by VjOOQ IC
90 KLBCTRIC SMBLTING AND RBFININO.
of an alloy by using a fused metallic cathode. But all experi-
ments made in this direction in the laboratories of the Aachen
Technical High School have so far given negative results.
Warren's FrocesB. — Warren* has, in the case of the bromide,
confirmed the efficacy of the method given above as applicable to
the haloid salts in general. He used for -the purpose about S>
kilogrammes (6 '6 lbs.) of beryl, purified by crushing and washing;
the powdered substance was then mixed with about four times
its weight of soda and fused in a furnace worked with a blast.
After cooling, the melt was dissolved with the aid of super-
heated steam, decomposed by means of hydrochloric acid, and
finally evaporated to dryness in order to render the silica in-
soluble. The solution separated from this residue was then
precipitated with soda, after separating out the iron and
chromium. Aluminium and beryllium oxide were next ex-
tracted from the moist residue by treatment by means of
sulphur dioxide; and the alumina was deposited from the
resulting solution as a granular precipitate on boiling. The
filtrate from this precipitate, on the addition of anmionium
carbonate, yielded beryllium carbonate, which was then mixed
with lamp black and heated in the absence of air, and treated
with bromine in a fireclay retort at a red heat. The metal was
deposited from the distilled bromide by a current of 8 amperes
at 12 volts. No further information was given in the original
paper as to the method of conducting the electrolysis, as to the
current-density, or as to the apparatus used. The most varied
objects were made from the metal.
REDUCTION OF BERYLLIUM OXIDE.
Eleotro-thermio Method of Beduotion. — The reduction of
the oxide by electro-thermal mecuis has been accomplished in
various ways. In the first German edition of this book the
author stated that all metallic oxides were capable of reduction
by electrically heated carbon, but also specially pointed out that,
although at tibat time the possibility of reducing beryllium oxide
was recognised, pure metal could not be obtained by this experi-
ment.
Liebmann,f in his patent for the production of beryllium
alloys, by a method fundamentally similar to the Cowles
process for producing aluminium alloys, was the first to con-
firm the reducibility of beryllium oxide by electrically heated
carbon. He heated to a white heat (necessarily, as the patent
specification states, by electrical means) a mixture of the oxide
with a reducing agent and the metal, or oxide of the metal, to
be alloyed, together with the requisite amount of charcoal.
* Chemical Ktwfi, 1895, vol. Ixxii., p. 311.
t German Patent 94,102.
Digitized by LjOOQ IC
BEBTLLIUM. 91
Iiebeau'8 XSzperiments. — Lebeau's further experiments^ as
to the treatment of beryl confirm the author's observations. If
the mineral be heated for a sufficiently long time in the carbon
crucible of the electric furnace, with a current of 950 amperes
and 45 volts, the mineral loses the greater part of its silica by
volatilisation, and becomes converted into a mixture of the car-
bides of aluminium and beryllium, together with ferro-silioon
and silicon carbide, from which dilute acids extract the aluminium
and beryllium, and dilute hydrofluoric acid only the latter metal.
If an intimate mixture of 100 kilogrammes [2 cwts.1 of beryl
and 50 kilogi'ammes [1 cwt.] of good powdered coke be heated
for an hour, by the application of a current of 1,500 amperes,
two layers are formed. The upper layer is made up of varying
quantities of silica, alumina, and beryllia ; it is not attacked by
acids, and is, therefore, useless for further treatment. The lower
layer, on the other hand, consists for the most part of crystallised
silicon, which may be obtained almost pure by treatment, first
with hydrofluoric acid and then with sulphuric acid. A method
is, therefore, available for the production of silicon on an indus-
trial scale, by the reduction of natural silicates by the agency of
carbon in the electric furnace.
Calcium carbide is found to be a convenient reducing agent.
On heating for an hour and a-half a mixture of equal parts of
finely-crushed beryl and coarsely-powdered calcium carbide, using
a current of 1,500 amperes for the purpose, a homogeneous, fused,
greyish-green mass was obtained, consisting chiefly of a mixture
of the aluminium and beryllium carbides. This material on
exposure to moist air resolved itself, owing to the oxidation
of the carbides, into a powder composed of the corresponding
hydroxides, from which from 90 to 95 per cent, of the beryllium
contained in the beryl could, without difficulty, be extracted by
the agency of hydrofluoric or sulphuric acid.
The ol)servation of Liebmann above referred to, which, in
view of the matter published in the first German edition of this
book and of ihe well-known Cowles process for the production
of aluminium alloys, contained nothing specially original, was
confirmed by Lebeau.t He found that on heating a mixture of
beryllia, copper oxide, and reducing carbon for five minutes^
with a current of 900 amperes and 45 volts, he obtained a brittle,
metallic, fused mass, with a rose-red fracture ; it consisted of a
pale yellow, or silvery-white alloy, that melted at the temperature
of a Ferrot furnace, and a red crystalline double-oxide of beryl-
lium and copper. The formation of the oxide can be avoided by
using a large excess of carbon and longer heating ; but in this
case most of the beryllium will become carburised, and the
greater part of tiie copper will be volatilised. It is, therefore,
*Compte8 Bendus, 1898, vol. cxxvi., p. 1203,
t/6id., 1897, vol. cxxv., p. 1172.
Digitized by VjOOQ IC
92 ELBCTBIO SMELTING AND REFINING.
better to remove the oxide by re-melting the crude product in
a Perrot furnace. The alloy thus purified contains from 5 to
10 per cent, of beryllium, and varies in colour, accordingly, from
yellow to nearly white. The yellow alloy, containing 5 per cent,
of beryUium, may be filed, is susceptible of polish, and may be
hammered either hot or in the cold; it withstands the action of
the air, and on melting with more copper affords alloys con-
taining less beryllium, among which that containing 1*32 per
cent, of that metal is gold-yeUow in colour.
Although, as will be understood from the foregoing, it is
possible to obtain beryllium without great difficulty, the small
quantity of the beryUium oxide contained in the raw material
available, and the necessity for separating the beryl from large
quantities of silica and alumina, render the process so costly that
it has hitherto found no application in the arts, notwithsttuiding
. that the properties of the metal, and especially its extraordinarily
^ow specific gravity, would ensure a good demand for it.
Digitized by VjOOQ IC
ALUMINIUM. 93
PART II.— THE EARTH METALS.
CHAPTER I.
ALUMINIUM.
Ocourrenoe of Aluminitun in Nature. — Aluminium is
found in nature onlj in the combined state. Of the com-
pounds that occur naturally, the oxide forms the basis of
corundum, sapphire, and emery; the hydroxide is present in
dlaspore, bauxite, and hydrargilllte ; and the salts are thua
represented : — The fluoride by cryolite ; the sulphate by the
alums, alunite, and alum shale; the silicates by the felspars
and their products of decomposition, the clays (kaolin, dec.).
Among these minerals, corundum and sapphire are used as
precious stones, emery for polishing and grinding, and felspar
and clay for the manufacture of building-stones, firebricks,
pottery, earthenware, stoneware, and porcelain. Chemical pro-
cesses of a more or less simple kind are employed to utilise
alunite and alum shale in the manufacture of alum ; cryolite in
that of soda and pure aluminium hydroxide and oxide ; and the
hydroxide in producing the pure oxide and hydroxide ; whilst
corundum was used in the Cowles process for the direct pro-
duction of aluminium alloys. None of the minerals can be
utilised for the direct extraction of pure aluminium.
Properties of the Metal. — Many of the properties of
aluminium have only been determined by means of experi-
ments * made during the last few years. It appears, therefore,
to be desirable to describe these properties in greater detail
than in the case of the other metals, especially as many errors
haye crept into metallurgical and chemical text-books.
Aluminium {Al'"; atomic weight » 27 ; sp. gr. » 2*58, when
chemically pure, or 2*6 to 2*7 in the case of the purest specimens
obtainable in commerce) has a white colour and good lustre, and
shows a crystalline structure when broken. Its melting point
has now been determined by Le Ohatelier with comparative
certainty as 625'' 0. Its specific heat at O*" 0. is 0*2220;
between 0* and 100** 0., 0*2270 (mean); and at the melting
* Bee Aluminium, by J. W. Richards, 3rd ed., Philadelphia and London,
1896; and Hunt, in Joum. of Franklin Inst., 1897, vol. cxliv., pp. 81, 171»
Digitized by VorOO^ It!
94 BLECTRIC SMELTING AND REFINING.
point, 0*285. Aluminium eyolves on cooling from its melting
temperature, 158 calories when solid or 258 calories when
liquid, so that its latent heat of fusion is 100 calories. The
coefficient of linear expansion of solid aluminium is 0*0000231,
and the contraction of liquid aluminium on solidification is
0017. Its conductivity for heat (silver = 100) is 31*33.
Aluminium is so friable at 530* C. that it may then be
pulverised.* When rolled or hammered in the cold, aluminium
rapidly loses its ductility, but this property is restored to it if
it be annealed by heating it to from 400* to 500^ C, and then
cooling it slowly. The temperature best adapted for rolling
or hammering the metal is 100* to 150* 0., and for pressing,
500* 0. Up to a certain point, both the hardness and the
strength of the metal are considerably increased by cold
working (hammering or rolling). For example, the tensile
strength of aluminium in the cast condition is about 12 kilo-
grammes per square mm. [7*6 tons per square inch]; but by
cold rolling the metal, so that it is subjected to a moderate
reduction in sectional area, the tensile strength is 18 kilo-
grammes per square mm. [11*4 tons per square inch]; after so
reducing the area from 20 to 1, the tensile strength is 23*5
kilogrammes per square mm. [14*9 tons per square inch]; or
by a reduction from 80 to 1, it is 27*5 kilogrammes per square
mm. [17*7 tons per square inch].
The electrical conductivity of aluminium, according to
Richards t (copper «= 100), is as follows : —
For metal with 98*6 per cent. Al,
. 56
„ ,t 99*0 ,,
. . 69
„ ,, 99*6 „
. 61
,, ,, 99-76
. 63 to 64
„ „ 100-00 ,.
. 66 to 67
Aluminium alloys with most metals, and the resulting alloys
are treated of very fully in Richards' Aluminium, It combines
chemically with almost all metalloids, sometimes with consider-
able energy. In the cold, however, and even up to a good red
heat, it strongly resists the action of oxygen, but at that
temperature it combines with it, and in so aoing evolves much
heat.
Water and dilute organic acids have scarcely any action on
the metal in the cold; but the latter very slowly attack it
when boiling. Nitric acid is almost without action on alu-
minium, and sulphuric acid dissolves it slowly; but hydrochloric
acid and caustic soda bring it very rapidly into solution. It
precipitates most of the other metals from solutions of their
* Aluminium Industrie AktiengeseUscho^, English Patent 16,969, Aug.
t«/bum. Franklin Inet.f 1897, vol. czliii
Digitized by VjOOQ IC
ALUMINIUM. 95
salts, and reduces most of the oxides when fused, even those
of carbon, silicon, and boron, any excess of aluminium present
then uniting or alloying with the reduced element.
The abundance of the natural compounds of aluminium, the
valuable properties of the metal, and, finally, the difficulties,
which must not be under-estimated, connected with the success-
ful decomposition of the raw material that is present in such
abundance, sufficiently explain how it is that the number of
researches, of inventions (either actual or merely patented), and
of more or less timid sugs^estions, has become almost intermin-
able. For the better estimation of the value of the processes
which haye been employed in the reduction of aluminium, it
will be desirable to consider separately the different divisions or
groups into which, from the metallurgical point of view, they
may be classed, viz. : — Precipitation, Reduction, and Electrolytic
Processes.
Precipitation Frooesses of Extracting Aluminium. —
Having regard to the scope of this book, a short account of
the more prominent processes of this class must suffice, for
the precipitation of one metal by the addition of another is
an operation which belongs to pure chemistry rather than to
electro-metallurgy .
Oerstedt,* in the year 1824, was the first to attempt the
decomposition of aluminium chloride by means of potassium
amalgam, but evidently without result; for other skilled
experimenters, working according to his directions, failed to
obtain any aluminium. W6hler,t however, in 1827, was
successful in reducing the anhydrous chloride by means of
potassium. Later, Deville | obviated the difficulties connected
with the production and use of aluminium chloride by employ-
ing the double chloride of aluminium and sodium instead ; and
he further substituted the much cheaper metal sodium for the
potassium recommended by Wohler. The process was actually
conducted on these lines for thirty years in France (first at
Nanterre and later at Salindres), and for a time in England
also. In 1855 Rose§ proposed the substitution of the mineral
fluoride (cryolite) for the chloride; and in place of sodium, the
use of which haa been adhered to by Kose, Beketoff || employed
magnesium. But the principal interest at present centres in
Grabau's proce8s,1I which is excellent in all its details. In this
process, solutions of sulphate of alumina are first treated with
* Oeratedt, Overs, o. d, Danske Vidensk. Selsk, ForJtandl., &o., 1824-25.
t Wohler, Popg, Ann., 1827, vol. xi., and JAebig^e Ann., voL liii.
X Ann. de CfhwUe et de Physique, 1854, voL xliz. See also H. St. Claire
PeviHe, De rAlvmimum, Paris, 1859.
%Poag. Awn., 1855, voL xovi.
i] Jahreaberieht der Chemie, 1865.
IT German Patent 47,031. [English Patents 14,356, Oct. 21, 1887, and
15,539, Nov. 14, 1887.]
Digitized by VjOOQ IC
96 ELECTRIC SMELTING AND REFINING.
cryolite, in order to obtain the aluminium entirely in the form
of fluoride, according to the equation : —
AW8O4), + ALjFe- 6NaF = 2AljF. + SNajSO^,
The aluminium fluoride, which is insoluble in water, is filtered
off, washed, dried, and heated to an incipient red heat, and is
then at once charged into a cold vessel lined with pure cryolite.
The required quantity of dry sodium, in the form of a cube or
cylinder, is now placed upon the hot powder, and the vessel is
immediately covered up. The following reaction then occurs,
accompanied by a great evolution of heat ; but in other respects
it proceeds quite quietly : —
2AlaF. + 3Na, = Al^ + A1,F.. 6NaF.
The aluminium is afterwards found at the bottom of the
vessel, melted to a regulus, and covered with a slag of cryolite^
which has been completely fused owing to the heat of the
reaction. The latter bye-product is available for the production
of fresh quantities of aluminium fluoride. Of all the chemical
processes, this is the only one which, given a cheap method of
producing sodium, is ever likely to come into competition with
the electro-chemical methods. The metal obtained by this process
has the advantage of being unusually pure.
The Beduction Prooesses. — For a long time alumina was
held to be non-reducible ; and even to this day statements to
that effect are to be found in the chemical text-books. This
has probably resulted from an erroneous comprehension of the
fact that pure metallic aluminium in a form that would be
useful in the arts is never obtained by the direct reduction of
the oxide. In addition to this, the high temperature necessary
for reduction can only be (or, at least, is most advantageously)
obtained by the conversion of electrical energy into heat. This
circumstance has frequently given rise to the assumption that
the reduction of aluminium is either entirely or in part due to
electrolytic agency. But the production of the metal is really
due, as will be shown later, entirely to a reduction of the oxide
of aluminium by electrically lieated carbon.
The idea of heating a material with a high electrical resistance
by a powerful current, and of placing this resistance in the most
intimate contact with the mixture to be reduced (the material
being distributed uniformly through the whole mass and be-
coming, ultimately, a necessary constituent of such a mixture)
unquestionably deserves the most careful attention. It is by no
means a new discovery, and if there be occasion for surprise, it
is less on account of the results that have followed the applicit-
tion of this principle than because these results have been sa
tardily effected. For the Philosophical Transactions* of the
♦ PhU, Trans., 1816, vol. cv., p. 871.
Digitized by LjOOQ IC
ALUMINIUM. 97
Royal Society of so early a date as 1815 contain the first record
of the metallurgical application of the heat generated by an
electric current in a resistance : — " Pepys . . . bent a wire
of pure soft iron so as to form an angle in the middle, in which
part he divided it longitudinally by a fine saw. In the opening
so formed he placed diamond powder, securing it in its situation
by two finer wires, laid above and below it, and kept from
shifting by another small wire bound firmly and closely round
them. All the wires were of pure soft iron, and the part con-
taining the diamond- powder was enveloped by thin leaves of talc.
Thus arranged, the apparatus was placed in the electrical circuit,
when it soon became red hot, and was kept so for six minutes.
. . . On opening the wire, Mr. Pepys found that the whole
of the diamond had disappeared, . . . and all that part (of
the iron) which had been in contact with the diamond was
converted into perfect blistered steel. A portion of it being
heated red and plunged into water became so hard as to resist
the file, and to scratch glass."
Depretz* first described a small apparatus, which is still very
suitable for many experiments : this consisted of a tube of sugar
charcoal rather more than \ in. wide, and almost ^j^ in. long. It
was closed by two plugs of the same material, and, being placed
in the circuit of a powerful electric current, became heated with
its contents to a white heat.
Even if the two arrangements just described have no direct
reference to the reduction of aluminium, they are, nevertheless,
worthy of consideration in connection with certain patent speci-
fications of a later date. The application of this principle to the
production of aluminium was first proposed by Monckton f in a
patent which he took out in England. A strong electric current
was to be passed through a reduction chamber charged with
alumina and carbon, in order that the mixture (by the agency of
the carbon) might be so highly heated as to lead to the reduction
of the earth metal by the carbon. At the time, this process was
not practicable, commercially speaking, for the production of
aluminium, even if it could have produced a sufficiently pure
metal, because no adequately cheap source of electrical energy
was available until 1867, when the dynamo was invented. But,
in addition to this, the product of reduction obtained in this way
could not have paid for the energy expended ; for the absorption
of carbon I could not be avoided in this process ; a certain loss of
metal by evaporation would be inevitable ; and, finally, the pro-
duct of the reduction would be a gray, brittle, crumbling mass,
with difficulty fusible, but just fritted together, and containing,
* Comptea Bendua, 1849, vol. zxix.
t Eng&sh Patent 264, 1862.
t See the chapter on the Reduction and Properties of the Carbide at the
end of this book.
7
Digitized by VjOOQ IC
98 BLECTRIC SMELTING AND BBFININO.
besides alnminium, aluminium carbide, carbon, and other im-
purities which must inevitably be derived from the carbon used
for reduction. Underlying the process, however, there was a
perfectly sound principle, which has found, and will find again,
many valuable applications in its proper sphere in metallurgy.
The Cowles Prooess. — After a long interval, the Brothers
Oowles introduced (in 1884) a process, by which there was made
available for use a number of alloys, of which the valuable pro-
perties had been in part known for a long time. Considering
that the impossibility of producing a sufficiently pure metal by
the direct reduction of alumina with carbon has been proved,
the invention of Oowles must be regarded as particularly fortu-
nate, for it consisted in alloying the aluminium, at the very =
moment of its production, with another metal, and so avoiding
the absorption of carbon, or at least reducing it to harmless pro-
portions. But without any wish to begrudge to the inventors
the recognition that is commonly awarded them, it must yet be
noted that their patent specification covered too wide a field.
In the principal patent * it is stated that the invention refers to
a class of furnace in which the electric current is employed as
the sole source of heat. Previous attempts had been made, it
was said, to reduce ores and to conduct other metallurgical
operations with the aid of the electric arc ; but this invention
consisted chiefly in the use of a granulated material of high
resistance, or low electrical conductivity, so connected in an
electric circuit that it formed an uninterrupted part of the same.
In consequence of its high resistance, it became red hot, and
afforded all the heat required. The substance to be reduced
was mixed with this granular material, and thus absorbed the
heat at the very place of its production.
The conversion of iron into steel must undoubtedly be recog-
nised as a metallurgical operation. If then Pepys, in the year
1815, as shown on p. 97, accomplished such an operation by
using a constituent part of the mixture as a resistance in an
electric circuit, and thus brought it to a red heat by the applica-
tion of a powerful current, while the other constituent of the
mixture was in direct contact with it, he was undoubtedly
working in exact agreement with the specification of the Cowlea
patent, published in America seventy years later. It was by an
application of exactly the same principle that Depretz (p. 97),
in the year 1849, placed a diamond in a tube of sugar-charcoal,
and included this in a powerful electric circuit, with the object
of ascertaining the behaviour of the gem in a neutral atmosphere
at high temperature. Finally, it might well be claimed for
Monckton that he anticipated the Cowles patent as far as it
refers to the reduction of alumina, by his specification published
in 1862 (p. 97).
♦U.S.A Patent 319,795.
Digitized by LjOOQ IC
ALUMINIUM.
99
These circumstances, however, can in no way be said to
detract from the service which the Brothers Oowles have done
in introducing so simple a process for the manufacture of
aluminium alloys.
Instead of entering minutely into the details of the various
forms of furnace which are specified in the numerous Cowles
patents, it will suffice here to describe a smelting plant on this
system, conoeming which detailed accounts have been published.
This plant, which was put down several years ago in the factory
of the "Cowles Syndicate Company," has been for some time
out of operation.* A 400 H.P. Crompton's dynamo gave a
Fig. 63. — Cowles' plant for manufacture of aluminium alloys.
current of 60 volts x 5,000 to 6,000 amperes. The smelting
furnaces were built of fireclay, and were rectangular in cross-
section. They were arranged side by side in a long row, but
were only put in operation one at a time, while the others were
in Tarious stages of cooling, emptying, or charging. The current
was conveyed to and from the battery of furnaces by two stout
copper bars running horizontally along the whole length of the
plant, and at a convenient height above it, one of these being
above the fronts, the other over the backs of the smelting
chambers. Running upon each of these rails was a copper
clamp mounted on wheels, with flexible cables of copper wire
*Indu8trie8, 1888, vol, cxv., p. 237.
Digitized by VjOOQIC
100
ELECTRIC SMELTING AND REFININO.
attached to it; the other ends of the cables were also held
together by clamps. A suitable opening in each of the lower
clamps permitted the latter to be hung upon copper bars, corre-
sponding in shape to the aperture, by which means electrical
connection could be made with the electrodes. Every electrode
consisted of a bundle of from 7 to 9 carbon rods, each 2^ in. in
diameter, around which a cylindrical head piece was cast, of
iron (if ferro-aluminium were being produced in the furnace), or
of copper (if aluminium-bronze were under manufacture). Into
Fig. 64. — The Cowles furnace (longitudinal section).
Fig. 65. — The Cowles furnace (cross-section).
the midst of this head piece was cast one of the copper bars
referred to above. The electrodes were introduced into the
smelting chamber through suitably laid cast-iron tubes in
opposite walls of the furnace. By adjusting a simple screw, the
electrodes could be moved backwards and forwards, as might be
necessary for the regulation of the current. In preparing the
furnace for use, the bottom was first covered with a layer of
limed wood-charcoal, and then the electrodes were introduced ;
a frame of sheet iron was next placed in the furnace, and the
Digitized by V^jOO^ It^
ALUMINIUM. 101
space within it was charged with a mixture of ore, metal, and
charcoal, whilst that between the frame and the walls of the
chamber was filled with charcoal only, and the frame was then
withdrawn. Some fragments of retort carbon were thrown into
the furnace to form a bridge for the current, and, finally, the
empty space still left was filled up with charcoal, and the whole
was surmounted by a cast-iron cover. In the centre of this
cover there was an opening for the escape of gases generated
during the smelting operation, the gases were lit, and the
products of combustion were led through a tube into a furnace
in which any alumina that they might contain should be
deposited. The melted alloy accumulated on the hearth of
the furnace, and was run off through a taphole placed there for
the purpose. The slag, which consisted of a very intimate
mixture of alloy and charcoal, was powdered and washed to
separate the charcoal, the residual alloy being introduced into
the furnace with a fresh charge.
In the furnace above described, fix>m 15 to 20 cwts. of ferro-
aluminium or aluminium-bronze, containing 15 to 17 per cent,
of aluminium could be produced daily. The bronze was re-
melted with the requisite proportion of copper to yield the
alloys required for the market^ containing respectively 1*25, 2*5,
5, 7 '5, and 10 per cent, of aluminium, and was cast into ingots
weighing from 10 to 12 lbs. each. The average electrical power
required amounted to 50 horse-power hours per kg. [which is
nearly equivalent to 23 H.P. hours per lb.] of aluminium. A
genend view of a smelting-house is shown in Fig. 63, whilst Fig.
64 gives a longitudinal, and Fig. 65 a cross-section of a single
furnace. In these figures EE are electrodes, consisting of
9 carbon rods, each about 1^ ins. thick, around which are cast
the cylindrical metal blocks, M. At the end of these blocks,
opposite to the carbons, the copper rods, K, are introduced,
llie whole arrangement slides in the tube, R, and may be moved
in either direction with the aid of the screw, S, and the nut,
which forms part of the collar, Z, attached to K. The current
is conveyed by the copper wire cables, L, which are gripped by
the copper clamp, Y ; and this in turn is rested on the conical
«nd of the rod, K, which projects beyond the mouth of the
tube, R. The cast-iron guide-block, F, serves to keep the rod^
K, in place.
Theory of the Cowles Process. — It is evident from an
inspection of Fig. 64, which shows the usual arrangement of the
carbon rods within the furnace during the operation of smelting,
that these rods, so often falsely termed electrodes in descriptions
of the plant, are really only resistances introduced into the
circuit of a powerful current. It is they which are first heated,
and which then impart their heat to the surrounding charge.
By degrees, as the charge around the inner extremities of the
Digitized by VorOO^ It!
102 ELECTRIC SMELTING AND REFINING.
carbon rods becomes heated to incandescence, the rods are with-
drawn from one another little by little by means of the screw, S.
Thus, to an increasing extent, as the heating extends, the
particles of carbon in the mixture form part of the circuit, and
so also act as resistances, whilst the rods, E, become partly
burned by the oxygen of the metallic oxides.
Hampe* refuses to admit that the heat generated by the
current can alone suffice to bring about the reduction of alumina
by carbon, because many attempts to reduce alumina by carbon
in the presence of copper or copper oxide have given negative
results, even though the high temperature of a Devil le's furnace
was employed. These experiments have led him to the
following conclusions : —
"If a temperature nearly equal to the melting point of
quartz be insufficient to enable carbon to reduce alumina in the
presence of copper, it is certain that the temperature of the
electric furnace, although it is much higher than this, will also
fail. I believe, therefore, that the production of aluminium-
bronze by the Cowles process is dependent far less upon a
(chemical) dissociation of alumina in the electric furnace than
upon the electrolytic decomposition of the alumina by the
current passing through it, the earth being brought into a state
of fusion by the electric arc. According to my view, then, the
action of the electric arc is primarily electro-thermic, but it
is also distinctly electrolytic."
The author cannot agree with this view, but believes, on the
contrary, as a result of his observations, that the decomposition
of the alumina in the Cowles furnace is attributable 8clely to the
influence of the heat generated by the current in the resistance tJuU
is introduced into the circuit ; and, further, that it is not neces-
sary to employ the higher temperature obtainable by the agency
of the electric current in order to bring about the desired result.
It is not beyond the range of possibility that even the tempera-
ture of a Deville's furnace should suffice for the initiation and
completion of the reaction. It is not possible, however, with
such an apparatus to fulfil all the conditions that obtain in
the electric furnace. For if a charge be placed in one or, as in
Hampe's experiment, two crucibles, and be then submitted to
the very high temperature of one of the best of these ordinary
furnaces, the heat of the surrounding fuel will be imparted but
slowly to the mixture enclosed in the crucibles. The tempera-
ture of the charge, therefore, rises quite gradually to the melting
point of copper or to the reduction temperature of copper oxide,
and long before the alumina and carbon shall have attained the
temperature at which aluminium could be produced, the copper
will have liquated out more or less from the mixture. The
evolution of heat from the union of copper with aluminium, that
♦ Chemiber Zeitung, 1888, p. 391.
Digitized by LjOOQ IC
ALUMINIUM. 103
Hampe had hoped would favour the reaction, cannot occur until
some free aluminium has been actually produced.
The conditions in the Oowles furnace are quite unlike those
just described. In this case, one of the charge constituents that
is to take part in the reaction — viz., the carbon required for the
reduction — is almost instantaneously raised to a temperature
which, as will be seen from an experiment shortly to be
mentioned, is amply sufficient for the reduction of the alumina.
Copper or copper oxide is not essential to the reaction, but is
added simply with the object of obtaining the aluminium in a
serviceable form as an alloy, since without this addition the
aluminium would combine with a portion of the excess carbon
to form a completely useless material, which no subsequent
refining operations would purify.
Thermic Reduction of Alumina by Carbon. — The author's
assertion may be proved by the following experiment (Fig. 66) : —
Fig. 66. — Arrangement to show direct reduction of alumina by carbon.
A thin carbon pencil, W, about ^ in. wide and 1| ins. long, is
fixed between two stout carbon rods, K, about 1 to 1| ins. in
diameter. This pencil is passed through a small paper bag, P,
about 1^ ins. in length, filled with an intimate mixture, M, of
alumina and carbon. The mixture is prepared by repeatedly
mixing and heating aluminium hydroxide with tar. The ends
of the paper cartridge are closed by small cork plates, and
the whole is packed in coarse charcoal powder. The electrical
connections are now made, and a current of 35 to 40 amperes is
passed through the arrangement for from two to three minutes.
After the cartridge has cooled sufficiently, the charcoal powder is
removed, and the carbon pencil will be found surrounded with a
Mtted mass, which, on examination, will prove to be aluminium
containing a considerable proportion of carbon. The possibility
of an electric arc having been formed is excluded by the condi-
tions of the experiment, and, in the absence of any break in the
conducting circuit within the charge mixture, the occurrence of
electrolytic decomposition is also negatived. It is unnecessary
Digitized by VjOOQIC
104 ELBCTRIC SMELTING AND REFINING.
to point out that an admixtare of copper or copper oxide with
the charge would have led to the production of aluminium-
bronze. With a simple arrangement of the above type it is,
therefore, easy to test the validity of the assertion previously
made, that given a sufficiently high temperatwre aU metallic oxide$
a/re capable of being reduced by carbon.
As above explained, it was necessary, in order to produce the
temperature required for the reduction of alumina, to pass a
current of 35 to 40 amperes through the J in. carbon pencil ;
this is equivalent to 5 to 6 amperes per sq. mm. [8,200 to 3,800
amps, per sq. in]. A current of 10 amperes per sq. mm. [6,500
amps, per sq. in.] will suffice for the most difficult case, and
should yield a temperature which, under these conditions, will
effect the fusion of any metal.
The electromotive force required to overcome the resistance
of such a carbon pencil, If ins. long, with the production of a
current ranging between the above-named limits of density,
would be 10 to 17 volts. With the aid of these data it is easy
to construct a resistance that shall be suitable to the current
employed in any given experiment. It must, however, be
pointed out that an attempt to work with a smaller current than
that named above will lead to great inconvenience and many
derangements of plant, because very thin carbon pencils are
easily broken, and may become burned through by the oxygen
of the metallic oxides within the first minute of use, even in
presence of an excess of carbon in the surrounding mixture.
Tests made with the most varied descriptions of carbon rods,
such as are used in arc-lamps (pencils ^ to ^ in. thick being
specially made for this purpose), have led to the adoption of the
following average numbers as sufficiently accurate for general
calculations: — An electromotive force of 0*3 to 0-4* volt is
necessary to drive a current of 1 ampere through 1 mm. of a
carbon pencil at the temperatures of these experiments, with
a current-density of 6 to 10 amperes per sq. mm. [4,000 to 6,500
amperes per sq. in ] of sectional area.
THE EIiECTBOLTTIC METHODS OF BEDUCING
AIiUMINIUM COMPOUNDS.
Electrolysis of Aqueous Solutions. — In order to trace-
without interruption the history of the development of the only
electrolytic methods that are technically practicable, it is
desirable to consider separately a whole class of impracticable
inventions, which aim at the electrolysis of aqueous solutions.
Just as the evidence advanced from different quarters cannot
be denied as to the possibility of precipitating aluminium in the
* See Borchers' Bau wid Betrieb eleHrischtr O^m, 1897.
Digitized by V^jOOQ IC
ALUMINIUM. 105
metallic condition from aqueous solutions, so it cannot be
admitted that, up to the present time, anyone has succeeded in
producing this metal by any practical process of the kind, either
pure, in merely just-recognisable traces, or alloyed, in any per-
manently stable form. Nevertheless, a short account will be
given of some of the proposals that have been patented or
otherwise brought before the public. The earliest records in
this subject consist of two English patent specifications.
Thomas and Tilley^ claimed to electrolyse an aqueous solu-
tion of freshly-precipitated aluminium hydroxide in potassium
cyanide.
Corbellit proposed the use of an electrolyte containing 2
parts of aluminium sulphate or alum, with 1 part of calcium
chloride or common salt, dissolved in 7 parts of water. The
anode was to be mercury (!), the cathode zinc.
The first August number of Dingler'a Journdt in 1854 con-
tained the following report, accompanied by a criticism that was
by no means flattering : —
** Reported Process for the Coating of Copper with Aluminium and
Silicon by Galvanic Means.— To obtain the aluminium, I boiled an excess
of dry hydrate of alumina in hydrochloric acid for one hour, then poured
off tne clear liquid, and added to it about one-sixth of its volume of
water ; in this mixture I placed an earthen porous vessel containing one
measure of sulphuric acid to twelve measures of water, with a piece of
amalgamated zinc plate in it. In the chloride of aluminium solution I
immersed a piece of copper of the same amotmt of immersed metallic
surface as that of the zinc, and connected it with the zinc by means of a
copper wire, and set it aside for several hours ; when, on examining it, I
found it coated with a lead-coloured deposit of alarainium, which, when
burnished, possessed the same degree of whiteness as platinum, and did
not appear to tarnish readily by immersion in cold water or in the
atmosphere, but was acted upon by sulphuric or nitric acid, either con-
centrated or dilute.
*' I found that if the apparatus was kept quite warm, and a copper plate
much smaller than the zinc plate was used, the deposit appeared in a very
short time, in several instances in less than half a minute. Also, I found
that if the chloride solution was not diluted with water, the deposit was
equally, if not more, rapid.
'* I have also succeeded in obtaining a auick deposit of aluminium in a
less pure state by dissolving ordinary pipeclay in boiling hydrochloric acid,
and using the supernatant clear solution undiluted with water in the place
of the before-mentioned liquid. A similar deposit of aluminium was also
obtained from a strong aqueous solution of acetate of alumina ; likewise
from a saturated aqueous solution of ordinary potash alum, but rather
slowly; with each of the solutions named, the deposit was hastened by
putting either one, two, or three small Smee's batteries in circuit.
*'To obtain the deposit of silicium, I dissolved monosilicate of pot^ish
(formed by fusing together 1 part of silica with 2^ parts of carbonate of
potash) in water, in the proportion of 40 grains to 1 oz. measure of water,
eding in like manner as with the alumina solutions, the process being
nedby interposing one pair of small Smee's battery in the circuit.
♦ English Patent 2,756, 1855.
t English Patent 507, 1858.
Digitized by VjOOQ IC
106 ELECTRIC SMBLTINQ AND REFINING.
With a very slow and feeble action of the battery, the colour of the
deposited metal was much whiter than that of aluminium, closely approxi-
mating to that of silver ; its other properties I have not yet had time
to examine" {George Gore* Birmingham).
The aathor remarks that the alaminium deposited upon the
copper plate dissolved even in dilute sulphuric or nitric acid;
but the aluminium produced by Deville exhibits a totally differ-
ent behaviour. The properties ascribed by Mr. Gore to his
metal are insufficient to identify it as aluminium; the same may
be said in regard to his silicon. The metallic coating, which
was found upon the copper plate in both instances^ is probably
nothing but zinc, reduced from the zinc sulphate that had formed
in the porous clay vessel at the expense of the dilute sulphuric
acid and the anode (zinc plate). {J, A^ickles.) t
Jeanson J electrolysed solutions of an aluminium salt of 1*15
to 1*16 specific gravity, at a temperature of 60® 0.
Haurd § recommended an aqueous solution of cryolite in
chloride of magnesium or manganese.
Bertram || proposed to separate the metal by a strong current
from solutions of aluminium-ammonium chloride.
J. Braunll (Berlin) proposed to obtain aluminium by the
electrolysis of an alum solution of 1*03 to 1-07 specific gravity
at the ordinary temperature. The sulphuric acid liberated
during the reaction was to be neutralised by alkali ; and the
separation of alumina was to be prevented by the addition of
a non- volatile organic acid.
According to a patent by Overbeck and Niewerth,** an
aqueous solution, either of organic salts of aluminium, of mix-
tures which afford such salts, or of aluminium sulphate with the
chlorides of other metals, is to be electrolysed.
The last part of the foregoing patent is also claimed by
Senet ft in his specification as his invention. He recommends
the moderate current of 6 to 7 volts and 4 amperes.
Next, A. Walter Jf appears with a patent, according to which
the production of aluminium is possible under the following
circumstances : — A solution of aluminium nitrate is to be so
decomposed with a small current-density^ but using the current
from a powerful dynamo^ and employing a platinised copper
plate, that aluminium may be deposited on the cathode in a
pulverulent form.
♦[P7aV. Mag., 1854 (March), p. 227.]
iJoum. de Pharmacie, 1854 (June), p. 476.
XAnnufil Record of Science and Industry , 1875. Vide Richards''
Aluminium, 1890.
§ U.S.A. Patent 2 >8,900, June 15, 1880. Vide Richards' Aluminium, 1890.
II Comptes JRenduM, 1876, vol. Ixxxiii., p. 854. T German Patent 28,760.
** English Patent 5,756, December 15, 1883.
ft Cosmoa len MondeM, 1885. (From Richards' Aluminium, 1890.)
XX German Patent 40,626.
Digitized by LjOOQ IC
ALUMINIUM. 107
About the time that this patent was published, in the year
1887, there appeared a recipe from the pen of Reinbold * for the
production of aluminium deposits on other metals : 50 parts of
alum were dissolved in 300 parts of water, 10 parts of aluminiuth
chloride were added, and the mixture was heated to 93* 0.
After cooling, 39 parts of potassium cyanide were mixed with
the solution. With an aluminium plate as anode, and, with a
weak current, a good polishable aluminium deposit was to be
obtained at the cathode. Even if this process had not been
given as a method for the production of aluminium, yet, had it
been workable, it would have afforded proof that the metal could
be precipitated from aqueous solution. This special case cannot,
however, be taken as controverting the statement made at the
beginning of this section.
R. de Monteglas,t in the first place, separated the iron electro-
lytically from a solution of aluminium chloride, and, after adding
lead, tin, or zinc oxide, claimed to deposit the aluminium in con-
junction with the added metal.
According to the specification of Falk and Schaag,| the
aluminium salts of non-volatile organic acids in aqueous solu-
tion are to be mixed with the cyanides of copper, gold, silver,
tin, or zinc, and after increasing the conductivity of the resulting
bath by the addition of alkaline nitrate or phosphate, the
corresponding alloys are to be separated from it by electrolytic
means.
Burghardt and Twining possess a number of patents which
have for their object the production of aluminium and its alloys
by the electrolysis of alkaline aluminates. After the addition
of cyanides, or of other compounds of an alkali and a metallic
oxide, aluminium, or one of its alloys (according to the nature
of the added salt), is to be deposited by the current at a tem-
perature of about 80" C.
Nahnsen and Pfleger'sg invention consists in a separation of
aluminium, aluminium alloys, and magnesium in coherent form,
which is to take place without the occurrence of secondary
reaction, if (in contrast to the methods usually employed) the
electrolyte be cooled, and any rise in the temperature of the
aqueous solution during electrolysis be guarded against by the
use of cooling agents. Thus, at a temperature of 40*" G. a separa-
tion of aluminium hydroxide occurs in considerable quantities,
but at 4** 0. all the aluminium is said to be obtained in the
metallic state.
Rietz and Herold|| consider a solution of aluminium, starch,
* Jetpeilers' Jojirnal, 1887.
t English Patent 10,607. Aug. 18, 1886.
t German Patent 48,708.
§ German Patent 46,753. [English Patent 8,552, May 23, 1889.]
ll German Patent 58,136.
Digitized by VjOOQ IC
108 BLECTRIC SMELTING AND REFINING.
and grape-sugar to be a suitable electrolyte for the deposition
of aluminium. After describing the manner of producing this
solution, they state that the aluminium separates out, but always
in a spongy form, by applying a strong current and using platinum
electrodes. The metal is to be subjected to strong pressure, and
then to be cast into bars, whilst the residual solution is to be
treated for the recovery of the grape-sugar.
The American* and Germanf journals, in 1892, recorded the
application of aluminium-plating to iron constructional work at
the works of the Tacony Iron and Metal Company, in Tacony,
Pennsylvania, but without describing the most important part
of the whole process from the electro-metallurgical point of
view — viz., the method of depositing the aluminium.
In 1897 Hunt,t who was at that time Director of the Pitts-
burg Reduction Company, published the first account of this
process, but he at the same time recorded its want of success.
Iron pillars, which had previously been coated thickly with
copper, were covered with an alloy of a little aluminium with a
very large proportion of tin, which unfortunately proved to be
durable for but a short time, the process used being as follows : —
The iron pillars, which were destined for use in public buildings
in Philadelphia, were first thoroughly cleansed, then coated
with copper to the thickness of 1*5 mm. [0*06 inch] in an
alkaline copper bath, and were afterwards transferred to an
electrolytic bath containing sodium stannate and sodium alu-
minate in the proportion of 25 : 75, potassium cyanide, and
water. The temperature of the bath was about 55" C. and
the current-density employed was 80 amperes per square metre
[7 '4 amperes per square foot]. Under these conditions an alloy
of 25 per cent, of aluminium and 75 per cent, of tin should have
been produced, but, as already stated, the coating formed did
not prove to be permanent.
It is unnecessary to criticise these processes, and other later
proposals made by €k>mers§ and Marino. ||
The opening statement can only be confirmed that there is
no hope whatever that any laurels are to be gained by an
attempt to obtain aluminium by the electrolysis of aqueous
solutions. The successful extraction of aluminium, in the state
of our present knowledge, can only be accomplished by the
electrolysis of fused compounds of aluminium.
Early Experiments in the Eleotrolysis of Fused Com*
pounds of AlnminiTim. — The first experiments in the pro-
duction of aluminium were concerned with the electrolytic
♦ Iron Age, 1892, Feb. 25, June 2.
t Stahl und Eiaen, 1892, Nos. 7 and 14.
X Joum. Franklin Inat., 1897, vol. cxliv., pp. 81 and 171.
§ English Patents 7,205, April 11, 1894 ; and 14,327, July 25, 1894.
II English Patent 20,354, Oct. 24, 1894.
Digitized by LjOOQ IC
ALUMINIUM. 109
deoomposition of the oxide, but were unsuccessful. They were
made by Davy* in the year 1807, after he had succeeded in
decomposing the alkaline hydroxides by the same process. The
alumina resisted the action of the current which Dayy had at
his command, but it is not improbable that later experiments f
afforded him an alloy of iron and aluminium. Working with
a Tessel charged with an atmosphere of hydrogen, he employed
the following arrangement: — A platinum plate was connected
up to the positive pole of a voltaic battery containing 1,000
couples. This plate carried a layer of alumina moistened with
water and kneaded closely together. Into the upper part of
this mass was introduced an iron wire which was joined to the
negative pole of the battery. The wire became instantaneously
heated to a white heat, and fused at the point of contact with
the alumina. The metallic mass, after cooling, was both whiter
and more brittle than iron. On treating it with acid a solution
was obtained from which alumina could be afterwards separated.
Bunsenl was the first to accomplish the electrolytic separation
of aluminium from its fused compounds, using for the purpose
the apparatus which he had designed for the reduction of
magnesium (p. 3), and adopting the readily fusible double
chloride of aluminium and sodium as electrolyte. But, as the
metal separated in pulverulent form at the low temperature of
the fusion, he gradually added common salt to the mixture
during the progress of the experiment until the temperature at
last was raised almost to the melting point of silver. After
cooling, the metal was found in large reguline spherical masses,
which could be melted into a regulus by projecting them into
fused common salt at a white heat.
Bunsen's account of these experiments was despatched to the
Editor of Poggendorff^s AnncUen on July 9, 1854 ; and only a
few weeks later, on August 14 of the same year, H. St. Claire-
Deville laid his thesis on metals, treating especially of alu-
minium, before the French Academy of Science.
Deyille's FroceBS. — As statements have crept into many
text-books to the effect that Deville intentionally left Bunsen's
process unnoticed, that portion of the thesis which has reference
to his experiments § with Bunsen's magnesium apparatus and
which adduces his reasons for modifying the arrangement is
here given in the form of a literal translation from the original
French. ||
" Tip to the present time it has appeared to me impossible to
obtain aluminium from aqueous solutions by means of the gal-
* Phil, Trans,, 1806, pp. 1 and 333.
t Phil, Trans., 1810, p. 16.
t Pogg. Ann,, 1854, vol. xcii.
%Ann, de Chimie ei de Phya,, 1854, vol. xliii., p. 27.
II [This translation is token from Deville's original paper.— Tbanslatob,}
Digitized by VjOOQ IC
110 ELECTRIC SMELTING AND REFINING.
vanic battery ; and I should even now believe in the absolute
impossibility of doing so if the brilliant experiments of Bunsen,
in the production of barium, chromium, and manganese, had not
shaken my convictions. However, I am compelled to say that
all the processes of this kind which have been published recently
in reference to the preparation of aluminium have given me
only negative results.
*^ Every one knows the beautiful process by means of which
Bansen has produced magnesium by decomposing magnesium
chloride with the aid of the galvanic battery. The illustrious
professor at Heidelberg has opened a way which may lead to
results that will be interesting from many points of view.
However, there can be no hope of applying the battery to the
direct decomposition of aluminium chloride, which is a substance
that does not fuse, but that volatilises at a low temperature; it is
necessary, therefore, to find a composition for the metallic bath
that shall involve the use of a fusible material from which
aluminium alone can be deposited by the electric current. I
have found such a substance in the double chloride of aluminium
and sodium, the production of which is a necessary feature of
the extraction of aluminium by sodium. This chloride, which is
fusible at about 185** C, and remains fixed at a sufficiently high
temperature, although it is volatile at temperatures above the
fusing point of aluminium, fulfils all the required conditions.
1 introduced this substance into a porcelain crucible, which
was imperfectly separated into two compartments by a plate
of biscuit porcelain, and decomposed it by means of a battery
of five elements, using carbon electrodes, the crucible being
heated, and the temperature being increased continually, in
order that the charge might be maintained in the fluid con-
dition as it became gradually less and less fusible; but the
fusing temperature of aluminium was not exceeded. Arrived
at this point, I stopped the experiment, and, after lifting out
the diaphragm and the electrodes, 1 heated the apparatus to
a bright red heat, and found at the bottom of the crucible a
regulus of aluminium, which was rolled and was exhibited to
the Academy at its meeting on March 20, 1854. It was
accompanied by a considerable quantity of carbon, which had
prevented a notable portion of the metal from uniting into a
single mass. This carbon resulted from the disintegration of
the very dense sample of retort carbon that served as electrode ;
and as a result of this action the positive electrode was entirely
eaten away in spite of its thickness, which was very consider-
able. This disposition of apparatus (as used by Bunsen for
magnesium) was not convenient in the case of aluminium ; and
the process to which I have been led, after many experiments,
is as follows : — The aluminium bath is prepared by weighing
2 parts of aluminium chloride and adding to it 1 part of marine
Digitized by V^jOO^ It!
ALUMIKIUM.
Ill
salt in the state of dry powder. The whole is mixed in a
porcelain crucible heated to about 200** C. Combination shortly
seu in with evolution of heat, and there results a very fluid
mixture, which is the bath used for the decomposition.
<*The apparatus [Fig. 67] consists of a glazed porcelain
crucible (P) which, as a measure of precaution, is placed within
a somewhat larger fireclay crucible (H); the whole is sur-
mounted by a crucible cover (D) pierced with a slot (L) through
which is passed a wide and stout sheet of platinum (K) to serve
as negative electrode, and with an aperture in which is tightly
fixed a well-dried porous cell (R). Within the latter is placed
a rod of retort carbon (A) as positive electrode. The bottom
of the ]>orouB cell should be kept at the distance of some
centimetres from that of the por-
celain crucible. The porcelain
crucible and the porous cell are
filled to the same level with the
fused aluminiiim-sodium chlor-
ide, and the apparatus is heated
after the manner described. The
electrodes are then introduced
and the current is passed through
the apparatus. Aluminium is
deposited with some sodium
chloride upon the platinum plate,
and chlorine together with some
aluminium chloride is disen-
gaged in the porous cell : fumes
are thus produced which are de-
stroyed by introducing dry and
powdered marine salt at inter-
vals into the porous cell. This
salt is transported to the negative
pole during the operation, along
with the aluminium. A small
number of elements (two are actually suflicient) are required
to decompose the chloride, which presents only a feeble resist-
ance to the electric current.
" The platinum plate is raised from time to tfme, as it becomes
sufficiently charged with metallic and saline deposit. It is
allowed to cool, the mass of salt is rapidly broken, and the plate
is replaced in the circuit The crude material detached from the
electrode is fused in a porcelain crucible enclosed within a fire-
clay crucible. After cooling, the mass is treated with water,
which dissolves a large quantity of sodium chloride ; and a grey
metallic powder is left, which is re-united into a regulus by
several successive fusions, adding double chloride of aluminium
and sodium, if necessary, during each fusion."
Fig. 67.— DeviUe's apparatus for
reducing aluminium.
Digitized by VjOOQ IC
112 ELECTRIC SMELTING AND REFINING.
Although the processes of Bunsen and Deville for extracting
aluminium could not have been carried into effect on a com-
mercial basis, owing to many difficulties that stood in the way,
their work must be said to have laid the foundation of our
present methods— viz., the electrolysis of fused aluminium com-
pounds.
Praotioal Obstacles to the Eleotrolysis of the Chlorides.
— Deyille's researches proved that the use of carbon cathodes
was not perfectly adapted to aluminium reduction. Experience
has shown that carbon rods or crucibles, when employed as
cathodes, appear almost to melt, so great is the disintegration
produced by the metal separated in their pores. An insur-
mountable obstacle to the practical use of this plant is found
in the chemical action of the electrolyte and deposited metal
upon every material of which the apparatus might be con-
structed ; and no suitable substance has yet been found for the
manufacture of melting and electrolysing-vessels, which are to
be heated externally, and used with the haloid salts of aluminium.
Both the fused salt and the metal itself absorb so many impuri-
ties in this way, during the electrolysis, that the valuable pro-
perties of the reduced aluminium are destroyed. The ordinary
fireclay and plumbago crucibles contain silicates which become
reduced by contact with the separated metal, and the latter is
thus contaminated with silicon. With this property is asso-
ciated that of very slight resistance to the action of melted
haloid compounds (excluding the fluorides). Porcelain vessels
not only have the faults of the above crucibles, but are very
fragile and costly, whilst they are only to be had in small sizes.
Crucibles of compressed carbon are so porous, that they cannot
be heated externally when filled with a liquid substance, unless
they are protected by an outer coating of an impermeable
material. Finally, there are no metals, which could be practi-
cally applied, that are sufficiently refractory, and that are yet
neither liable to attack by the electrolyte, nor capable of alloying
with aluminium.
The production of pure aluminium by these methods, there-
fore, has failed, owing to the want of suitable materials for the
melting vessel.
The Teaohing of Deville's and Hansen's Experiments. —
Deville, in the treatise to which reference has been made,
brought forward a suggestion that has hitherto almost escaped
notice, but that deserves to be better known than it appears
to have been, by reason of the important principle that it
enunciates. He there describes the conditions under .which
metallic objects, especially those made of copper, may be ooated
with aluminium by means of his process. In this case, in order
to maintain a constant proportion of aluminium in the fused
aluminium-sodium chloride bath, he recommends the use of anodes
Digitized by V^jOO^ ICl!
ALUMINIUM. 113
of aluminium or of a compresMd mixture of alumina and carbon*
He thus described two most noteworthy principles, which have,
beeu repeatedly re-discovered and patented, viz. : —
1. The use of soluble anodes injused electrolytes ; and
2. The addition of aluminium to fused compounds of the metal
during electrolysis, by the agency of alumina.
The first patent which contained the same idea was taken out
in England under the name of Le Cfaatelier.f In this case it
is probable, however, that the patent was applied for with
Deville's full knowledge, because Deville frequently refers in
laudatory terms to Le Ghatelier as his collaborator. But, in
addition to the use of alumina-carbon anodes, this patent also
covers the use of a porous cell for the reception of the anodeSf
which tend to disintegrate during the electrolysis and so to
introduce impurities into the electrolyte. It is, however, need-
less to point out that porous cells could not possibly be employed
in the melted haloid salts of aluminium for the manufacture of
that metal. But since the use of alumina- carbon anodes is
accompanied by a tendency to disruption cUid hence by the'
introduction of impurities into the bath, and since the bath
itself has a very low electrical conductivity, it is evident that
Deville's invention cannot be applied in this form. At the
same time, the fact that the practical application of Bun sen's
and Deville's processes is not directly possible must not be
allowed for a moment to detract from the service which these
investigators have done in preparing the way for the modern
developments of their ideas. It is necessary to understand the
reasons why experiments have failed in order to attain success.
The work of Bunsen and Deville formed the basis upon which
has rested the success obtained in this field within the last few
▼ears, as may be shown by the following deductions that may
be drawn from the results (both positive and negative) of their
work, in reference to the properties of the raw materials, final
products, and substances used in the construction of the required
apparatus : —
1. Aluminium may be obtained by the electrolysis of fused
{anhydrous) aluminium compounds.
2. The eUuminium separated from the fused material may be
replaced by the use of cUumina, so that the process may be made
more continuous.
3. The addition of aluminium to the electrolyte is not to be
effected practically with the aid of anodes composed of alumina
and carbon, although
4. The use of soluble anodes for the electrolytic refining of
metals is of great importance.
•H. St. Claire-DeviUe, Dt r Aluminium (1869), p. 96.
t English Patent 1,214, 1861.
o
Digitized by VjOOQ IC
114 ELECTRIC SMELTING AND REPINING.
5. Cathodes made of carbon could not be employed tmder tlie
conditions of working at that time known (p. 113).
6. There is no material known that is suitable for mdting
vessels to be nsed for the electrolysis of aluminium compounds
when external heating is applied.
This, then, was the condition of afihirs at about the middle
of the nineteenth century. The last obstacles were removed,
almost thirty years later, by the solution of the difficulties
mentioned under the fifth and sixth heads. Before passing to
the methods at present in use, it will be advisable to mention
separately some of the proposals and experiments which, to
judge at least from their original descriptions, could not possibly
lead to the desired result.
Impraoticable Prooesses. — First among these is Gaudin's
process,* by which a melted mixture of cryolite and common
salt was to be decomposed by the current into aluminium and
fluorine. That of Kagenbuschf scarcely requires criticism:
clay, melted with suitable fluxes, was to be electrolysed with
the addition of zinc ; and the zinc was afterwards to be removed
from the alloy by distillation or by a refining process. The
process specified in Berthaut's patent} is practically identical
with that invented by Deville. Faure's apparatus (1880) for the
decomposition of aluminium chloride may also be passed over
without comment.
Graetzel's patent § describes a process that is well known but
quite impracticable. It is, however, referred to in many chemi-
cal text-books as being not only sound in principle, but as being
in actual operation. It must, therefore, be examined somewhat
more carefally than would be otherwise necessary. The inventor
proposes to electrolyse melted chlorides or fluorides in the
apparatus shown in Fig. 68. The melting vessel, s, is made of
porcelain, stoneware, or similar fire-resisting material, and is
protected from the direct action of the flame by a metal jacket.
Within the inner vessel is a cathode of metal, preferably of
aluminium ; whilst the anode is a carbon rod, K, enclosed in a
porcelain tube, G, provided with openings, g^ below and an
escape pipe, p, for chlorine, above. During electrolysis a current
of reducing gas is circulated through the fusion chamber, B, by
means of the tubes, 6^ and o^. In order to reduce the electro-
motive force required, as well as to maintain the strength of the
bath, which would otherwise become impoverished, there are
introduced into the inner vessel, G, besides the anode, but quite
independent of it, plates or rods, M, which are composed of
equivalent weights of alumina and carbon. The carbon is to
* Moniteur ScierUifique, vol. xi., p. 62. (See Richards' Aluminium, 1890.)
t English Patent 4,81 1, 1872. (See Richards* Aluminivm, 1890. )
::: EngUsh Patent 4,087, 1879.
§ German Patent 26,962.
Digitized by VjOOQ IC
ALUMINIUM.
115
combine with the oxygen of the oxide, whilst the metal of the
latter passes into combination in the bath. Unfortunately, the
carbon does not give the desired result, but only serves to
introduce impurities into the electrolyte. In order to obtain
the aluminium pure, it must be separated in the molten
condition ; but in that case it may well be asked what would
become of the aluminium cathode, the clay melting vessel, or, if
these survived, of the separated aluminium itself ? The intro-
duction of reducing gases is at the least superfluous in the
electrolytic deposition of aluminium. After the foregoing
remarks, there is but little of the invention left for criticism.
Lest, however, it should be said that those who were tempted
to repeat the experiments with this apparatus, were so unskilful
that they failed to recognise
its advantages, it should be
added, as an especially sig-
nificant fact, that Graetzel
himself in his capacity as
manager of the Hemelinger
Alumini'Wfnr und Magnesium
Fabriky did not use his own
process for obtaining alu-
minium, but that of Beketofif
(see p. 95), which consisted
in the reduction of cryolite
by means of magnesium. He
applied for a patent for his
process in various countries,
but the application was re-
fused, to the best of the
author^s knowledge, every-
where excepting in England.
The process brought forward by Boguski-Zdziarski,'^ although
not practically workable in the manner described, has yet some
interest in connection with certain later patents. Oryolite, or
other aluminium compound, is mixed with suitable fluxes, and
melted in a wrought-iron or plumbago crucible, heated in any
convenient furnace. At the bottom of the crucible is placed a
metal which is to be alloyed with aluminium ; and, during
electrolysis, this metal forms the cathode, while the anode is a
carbon rod dipping into the fused bath. The fluxes actually
forming the electrolyte must contain substances which can
readily combine with the fluorine separated at the anode, whilst
they should also be as rich as possible in aluminium. Mixtures
answering to these requirements may be obtained by fusing
cryolite or other aluminium compounds with carbonate of soda
or potash, or with other carbonates behaving in a similar way
* English Patent 3,090, Feb. 11, 1884.
Fig. 68. — GraetzeFs alumininm
electrolysing- vesseL
Digitized by LjOOQIC
116 ELECTRIC SMELTING AND REFINING.
in respect to flaorine. After some general remarks upon the
expenditure and the action of the current, and the injuriousness
of silica when present in the metal mixture, it is further
remarked : — " When a mixture of an oxide of aluminium
(AI2O3) is employed, the flux must be used in a quantity
sufficient to give an easily fusible combination. In this case the
flux should consist of carbonate of potash, or soda, or both."
The ideas expressed in this patent specification undoubtedly
contain more matter for consideration than do those of most
other inventions, which are, for practical reasons, unworkable.
Farmer* proposes (as an alternative amongst other aluminium
compounds) to mdt and electrolyse even cUuminium chloride in
open vessels. The melting-vessel is to serve also as cathode } and
the patentee thus revives in the year 1885 an idea which had
been well known so long ago as 1808.
Grousilliers,t in searching for a patent, has not forgotten that
aluminium chloride is easily vaporised under ordinary condi-
tions, and therefore proposes to accomplish the electrolysis of
this salt in closed vessels under pressure — a method of work
which is obviously impracticable.
At first sight (but only thus) Grabau's patent specification X
would appear, by the use of cooled electrode cells, to afford a
means of preventing that absorption of impurities by aluminium,
which is so troublesome when fused salts are electrolysed. In
explaining his process, Grabau gives the following example: — In
the electrolysis of a fused bath of cryolite and sodium chloride
it is known that chlorine is evolved at the positive pole, and
aluminium in the liquid condition at the negative. But since
melted cryolite is capable of attacking every known fire-resisting
and non-conducting material, this electrolytic process can only
be applied when, as is rendered possible in the present process,
the affected parts of the apparatus are protected from the action
of the fused salts or the separated constituents of the bath, by
the use of an insulating and resisting shield. In Fig. 69 is
shown an arrangement suitable for use in the process under
description. The iron melting vessel, A, is heated by external
firing until the bath is thoroughly fluid ; and the level of the
latter should then be at XX. The hollow metallic cylinder, B,
is made with double walls, through which, for cooling purposes,
a fluid such as water or air may be circulated, r being the inlet
and r^ the outlet f)ipe Another double- walled vessel, in the
shape of a trough, C, is employed as a reservoir for the reception
of the melted aluminium which separates. The liquid or gas to
be used in cooling this vessel enters the space between the walla
by r* and leaves it by r^. In consequence of the cooling action
♦U.S.A. Patent 315.266.
t German Patent 34,407. [English Patent 8,478, July 14, 1885.]
t German Patent 45,012.
Digitized by LjOOQ IC
ALUMINIUM.
117
produced by this circulation, a portion of the melted salt
solidifies over the whole of the cooled surface of the cell,
B, of the collecting vessel, C, and of the tubes, and forms at
these places a non-conducting crust, which is not capable of
being attacked either by the &sed salts or by aluminium.
Grabau starts from the assumption that only those parts of
an electrolyte vessel containing fused cryolite, which come in
contact with the separated elements, can lead to the introduction
of impurities into the bath and metal. If this were true, he
would certainly almost have afforded a solution to the difficult
problem of extracting pure aluminium electrolytically. He can
only be said in that case to have almost solved the question,
Fig. 69. — Grabau's cooled cell appai-atus.
because he does not state in his specification the nature of
the cathode, with which the separated particles of aluminium
will have to come into the most intimate contact. Further, it
appears to be doubtful whether, by the use of the anode cell aa
it is shown in Fig. 69, the chlorine would be efficiently removed.
The author's experience is that a large proportion of the chlorine
which rises from the anode, tends to float on the surface of the
fused salt, and to approach the walls of the vessel, and there, by
contact with iron, to introduce impurities into the bath, and
hence into the resulting aluminium. But putting these imper-
fections aside, the author has never succeeded in electrolysing
mixtures of aluminium salts or double salts, containing fluorine,
in iron vessels without both the bath and the metal taking up
Digitized by V^jOO^ It^
118 ELECTRIC 611SLTINO AND BBFININO.
much iron. The want of dnrability of iron yessela when used
for these processes is in itself both a proof of this assertion and a
ground of objection to snch processes.
Lossier^ alone has up to the present time, patented the
electrolytic separation of pure aluminium from aluminium
silicates.
Richards t describes an unpatented process of Rogers, ^ which
aims at electrolysing melted cryolite with a cathode of molten
lead. Here an alloy of lead and sodium is used initially, and
this separates aluminium at the expense of the sodium. The
American Aluminium Company, of Milwaukee, which was
founded in 1887 to apply this process, appears, however, to have
contented itself with the erection of a small experimental plant.
A. Winkler, § of Gorlitz, has not stated the nature of the
material of which his electrolysis-vessels were made, but as the
phosphates and borates of alumina were used as electrolytes, he
must have met with even less success than did his numerous
predecessors. The separation of pure aluminium from a bath
of ftised borates — even from aluminium borate itself — could
scarcely be proposed seriously by anyone who has ever seen an
account of Wohler's work with boron.
Patents granted to Feldmann recommend the use as electro-
lyte first of mixtures of the double fluoride of aluminium and
barium, strontium, calcium, magnesium, or zinc, with a chloride
of one of the latter metals j|| and afterwards of the haloid salts
of aluminium with the oxides of more electro-positive metals. IT
The latter process has also been patented by Cowles.**
Daniell ft has proposed an unlikely apparatus for the electro-
lysis of fused aluminium-sodium chloride ; and even combines
it with a plant for the production of aluminium chloride by
Wohler's method.
Diehl 1 1 patents the electrolysis of the fluoride, ALF^ . NaF^
prepared by a special method, using sodium chloride in the
fusion and a carbon electrode. Having recognised the useless*
ness of such cathodes (p. 112) he substitutes for them, at a later
date, others consisting of ferro-aluminium (AlgFe), which are to
take up aluminium during electrolysis until they contain the
equivalent of Alj,Fe. He works with a current-density of 1 3 to 20
amperes per square inch. The excess of aluminium contained in
the richer alloy is removed by heating, and the residuf^l AljFe
is again used as cathode.
* German Patent 31,089. t Richards' Aluminivm, 2nd Ed., 1890.
X Proceedings of the Wiscoimn NaturcU History Society, April, 1889.
§ German Patent 45,824.
II German Patent 49,915, 1887. [English Patent 12,575, Sept. 16, 1887.]
IT German Patent 50,370, 1889. •* English Patent, 11,601, 1890.
tt German Patent 50,054, 1889. [English Patent 4,169, March 9, 1889.}
tt German Patents 59.406, 69,447; and English Patent 813, Jan. 16,
4889.
Digitized by LjOOQ IC
ALUMINIUM. 119
Other inventions, the details of which were either previoasdj
known or are impracticable, have been made more recently by
Berg,* Dixon,t Graetzel,^ I>iehl9§ Faure,|i Ca8e,ir Gooch and
Waldo,** Roger,tt and others.
Successful Bedaction Processes. — Some twenty years ago
attempts were made to produce the heat necessary for the fusion
of the aluminium compounds, which were to be electrolysed, by
the conversion of electrical energy within the melting vessel
itself; and, in this way, the key to the successful solution of the
problem has been found. The production of heat by the agency
of electricity was attempted in two ways : either by introducing
the electric arc into the melting chamber or the mass to be
heated, or by including the electrolyte as a resistance in the
electric circuit. Only the latter way has proved economically
successful ; but some^ attention may be devoted in the first case
to those processes and arrangements which have for their object
the decomposition of aluminium compounds in the zone of the
electric ara
Decomposition of AluTnlnium Compounds by the Eleo<i
trie Arc. — The electric arc serves even too well for the purpose
of aluminium extraction. Although the two required conditions
of heating and electrolysis are here available, yet the tempera-
ture is unnecessarily high for the end sought, even when a
small arc is employed; and, moreover, this excess of heat is
concentrated within a very small area. The use of the electric
arc for the extraction of aluminium must, therefore, be considered
extravagant.
The impulse to use this source of heat is traceable to the pub-
lication by Sir W. Siemens of an account of his electric smelting
furnace, of which a description will be given after reference has
been made to some of the earlier experiments and proposals.
The temperature obtainable in the electric arc seems to have
been first utilised in experiments made by Depretz. In a note
to the French Academy of Science, of December 17, 1849, J J he
describes the behaviour of a small retort of sugar charcoal (of
about ^V^nch diameter) within which the arc was formed with
the aid of a carbon point. The retort itself served as positive
electrode.
Johnson's Process. — A somewhat later invention for the
smelting of ores was patented by J. H. Johnson §§ in England on
March 22. 1853. The ore was mixed with carbon and placed in
the electric arc, which was caused to play between two large
* German Patent 56,913, 1889. [Cf, English Patent 2,002, Feb. 6, 1890.]
f English Patent 16,794, Oct. 24, 1889. tOerman Patent 58,600, 1890.
§ German Patent 62,.^53, 1891. il German Patent 62,907, 1892.
f U.S.A Patents 512,801-512,803, 1894.
•*U.S.A Patents 627,846-527,851, 528,365. 1895. [English Patent
20,615, Oct. 17, 1894.] ft German Patent 83.109, 1895.
tXComptes Rendm, 1849, vol. xxix. §§ English Patent 700. 1853.
Digitized by LjOOQ IC
120
ELECTRIC SMELTING AND REFINING.
electrodes ; the ore thus became fused and converted into metal
and slag. These two substances fell into a receptacle placed to
receive them, where, by means of a suitable furnace, they were
kept in the fluid condition until the metal had separated from
the slag (compare Gerard^Lescuyer, p. 125). The same specifica-
tion contained also a description of a somewhat different appa-
ratus. The two electrodes were arranged so as to include an
angle between them. The upper was hollow and was filled with
the ore to be reduced, which was to be gradually impelled for-
wards by a screw. In connection with this description, it is
interesting; to examine a sketch (Fig. 70) which is taken from an
English patent specifi-
cation of H. Cowles.*
In this figure, E E are
the electrodes, iuto the
upper of which the ore
is fed by means of a
funnel, and so passes to
the neighbourhood of
the arc, while the fused
products pass out be-
neath.
It is not impossible
that the Johnson named
in the above patent was
only the agent for the
real inventor, for the
first part of the above
<le8cription would suit
an apparatus which was
described by Pichou in
the year 1853, and is
considered by Andreolif
to be the first electric
smelting furnace; this
apparatus is shown in Fig. 71, which is reproduced from the
original sketch. Here, O O are carbon rods gripped by the
holders, A A, x x are the conducting wires, T T are screws to
advance the carbons into the furnace, M is masonry, N is the
smelting hearth with a grate placed beneath it, and L is the
chimney. In the then state of technology the practical use of
this invention was out of the question ; but the possibility of
applying it was never lost sight of
The Siemens Eleotrio Furnaoe. — Sir William Siemens'
electric furnace t for refractory materials showed in principle
little that was new. It was not proposed to use it for the
♦ Knglish Patent 4,664. 1887. t Industries, 1893.
t English Patent 2,110, 1S79.
Fig. 70. — Cowles' ore-smelting furnace.
Digitized by LjOOQ IC
ALUMINIUM.
121
extraction of aluminium. The whole arraugement of the details
of the furnace, however, forms so interesting an object of corn-
Fig. 71. — Pichou's smelting funiace.
an with the productions of the modern inventor, shortly to
described, that a description of the apparatus must not be
omitted here.
Fig. 72. — Siemens' electric furaace.
Siemens himself has described * the apparatus. A crucible, X
(Fig. 72), made of graphite or other very refractory inaterial, ig
* Ekktrolechnische Zeitschrift, 1880.
Digitized by LjOOQ IC
122 ELECTRIC BMBLTJNG AND BKFININO.
placed in a metallic case, H, the space between the two vessels
being filled with powdered charcoal or other bad conductor of
heat. A rod of iron or platinum, or of gas-carbon, such as is
used for electric light purposes, is passed through a hole in the
bottom of the crucible. The cover of the crucible is also bored
in order that the negative electrode may be passed through it.
For this electrode is used, if possible, a cylinder of compressed
carbon, which should be of uniform measurement; it is sus-
pended from the end, A, of a beam, A B, supported in the centre
by means of a strip of copper or other good conducting material^
whilst to the other end, B, of the beam, a hollow cylinder of soft
iron is so fastened that it can move freely in a vertical plane
within a coil of wire, S, which offers a total resistance of about
50 ohms. The magnetic force by which the hollow iron cylinder
is drawn into the solenoid-coil, is balanced by a counterpoise, G,
which may be moved freely along the beam, AB. The resistance
of the arc is determined, and, within the limits imposed by the
source of power is fixed, by the shifting of the sliding weight to
any required point. The one end of the solenoid-coil is con-
nected to the positive, the other to the negative, pole of the
electric arc. Hence an increased resistance of the arc leads the
negative electrode to sink to a greater depth within the crucible,
whilst a diminished resistance causes tlie counterpoise to press
the iron cylinder deeper in the coil, whereby the length of the
arc is increased until equilibrium is restored between the
opposing forces. This automatic regulation is of the highest
importance, for without it the temperature in the crucible would,
on the one hand, diminish, while, on the other hand, the sudden
diminution in the electric resistance of the materials during the
melting would produce not only a sudden increase in the resist-
ance of the arc, but probably also its extinction. It is essential
to the satisfactory performance of electric fusion, that the
substance to be melted should form the positive pole of the arc,
because, as is well known, heat is generated at this pole. This
arrangement is evidently suited for use with metals only, for
non-conducting earths or gases it is necessary to provide an
indestructible positive pole which is also capable of being melted,
and so of forming a layer of liquid at the bottom of the crucible.
The heat accumulates very rapidly. With the aid of a medium-
sized (36 " Webers" *) dynamo a crucible 8 ins. deep, set in non-
conducting material may be brought to a white heat in less
than a quarter of an hour, and 2 lbs. of steel may be melted in
it within half an hour from the commencement. To check the
consumption of the negative pole, the author [Siemens] uses
a water-cooled electrode, or a tube of copper through which
a stream of cold water is made to circulate. This consists of a
* [The Weber was the old electrical unit of quantity ; in the above place
it may be translated amperes. — Translator.]
' Digitized by
Google
ALUMINIUM.
123
simple copper cylinder, closed at the lower end, and containing
a caoutchouc tube which reaches nearly to the bottom, and
serves to convey into it a stream of water.
The adjoining sketch (Fig. 73) is taken from the English
patent specification.* Of the electrodes, A is made of carbon,
whilst B is a metal tube cooled by cold air or by water, as in
Fig. 72. Each electrode is guided by a pair of rollers, R, r.
A common crucible steel furnace uses from 2^ to 3 tons of the
best coke per ton of steel melted, a regenerative crucible furnace
only 1 ton, and a regenerative gas furnace when used in con-
nection with an open hearth, only about 12 cwts. The electric
furnace therefore approximates to the regenerative gas furnace
in respect of fuel economy. But the former has the following
advantages: — 1. The temperature attainable is theoretically
unlimit^. 2. The charge is melted in a completely neutral
atmosphere. 3. The process can be conducted without much
Fig. 73. — Siemens* modified electric furnace.
preliminary preparation, and under the direct supervision of the
experimenter. 4. Even when using the ordinary refractory
materials, the limit of temperature that may be practically
reached is very high, because in the electric furnace the sub-
stance which is being heated is hotter than the crucible, whilst
in ordinary melting processes the temperature of the crucible
must exceed that of its charge. Although the electric furnace
may not supplant the ordinary farnace, yet chemical reactions
of the most varied kind may be in future carried on with its
aid, and at temperatures hitherto unattainable. In course of
practical trials with this apparatus, it was found that 20 lbs.
of steel could be completely melted in an hour, and 9 lbs. of
platinum in a quarter of an hour, whilst copper which had been
packed in carbon dust lost more than 90 per cent, of its weight
by volatilisation.
The Kleiner-Piertz Eleotrio Furnace. — Although the fur-
English Patent 4,208, 1878.
Digitized by VjOOQ IC
124
ELECTRIC SMELTING AND REFINING.
naces just described were not expressly intended for the treatment
of aluminium compounds, they embodied the fundamental ideas
of the arc furnaces which have since been proposed and experi-
mentally tried for aluminium extraction, as may be seen, for
example, from the following account of the construction of the
Kleiner-Fiertz furnace,''^ in which aluminium fluoride was to be
treated for the separation of aluminium. The electrodes of the
apparatus (Figs. 74 and 75) dip into a vessel, B, lined with clay
Fig. 74. Fig. 75.
The Kleiner-Fiertz aluminium furnace.
and filled with cryolite and bauxite. The negative electrode, K,
is capable of adjustment in a vertical plane, and the position of
the positive pole is controlled by a weighted lever and a solenoid,
its movements being limited and checked by a piston, I, plung-
ing into a liquid, and placed above. It is difficult to say what
advantage this arrangement has over that of Siemens. The
practical success of the method, no doubt, remains far behind
that anticipated for it by the inventors ; but it is not possible
* German Patent 42,022, 1886. [Compare English Patents 8,531 , June 29,
1886, and 16,322, Nov. 24, 1886.]
Digitized by V^jOO^ li:!
ALUMINIUM.
125
to imagine continuous work with any material, which, although
comparatively easy to melt, has ample opportunity for re-solidi-
fying owing to the concentration of the heat of the arc at so
small a point in immediate proximity to cold material. An
interruption of the current would be much more frequently
necessitated by the freezing of the electrode into the charge,
than by the decomposition of a sufficient quantity of the double
fluoride in the bath.
The Qrabau Electric Furnace. — Grabau's process and
apparatus for melting or reduction by means of the electric
arc* have not removed the difficulties described in connec-
tion with the Kleiner-Fiertz plant ; indeed, they have the
appearance of exhibiting these disadvantages in an even higher
degree. The material to be melted is introduced beneath the
surface of the fused mass in the crucible by means of tubes,
and this mass forms one
pole for the passage of the
current. Various modifica-
tions are illustrated in the
patent specification which
allow the material under
reduction to be added con-
tinuously, either alone, or
associated with the metal
that is to be alloyed with it.
The alloying-metal at the
same time is made to conduct
the current. The fused mass
is kept at a constant height
by means of an overflow
notch. The illustration and
detailed description of this
apparatus need not be given here for the reasons already
indicated.
The Gterard-IieBCuyer Furnace. — The apparatus (Fig. 76)
of Gerard-Lescuy er t affords a very good illustration of the
furnace patented by Johnson in 1853. Instead of mixing the
ore with carbon before submitting it to the action of the arc,^
he moulds it into bars with the aid of pressure, and in admixture
with copper, alumina, carbon, and tar, and then uses these as
electrodes for the production of the arc. In other respects the
process is sufficiently described by Johnson's specification (p. 120).
The Willson Furnace. — References to Willson's I process*
of producing aluminium are frequently to be found in technical
literature. But, by the light of his patent specifications,
* German Patent 44,511, 1886. t German Patent 48,040, 1887.
±U S.A. Patents 430.463, June 17, 1890; 492,377, Feb. 21, 1893. English
Patents 4,767, March 17, 1891 ; 21,696, Nov. 28, 1892 ; 21,701, Nov. 28, 1892.
76. — The Gerard-Lescuyer
electric furnace.
Digitized by V^OOQ IC
126
ELECTRIC SHELTIKO AND REFININO.
aluminium could scarcely be reduced profitably, because, like
Kleiner- Fiertz and others, he employs the electric arc for heating
and electrolysis, which is both costly and extravagant. His first
furnace was, at least, so far unlike that of Siemens' that he had
introduced an arrangement by which reducing gases (or carbon
powder and the like) might be injected into the arc,**^ in order
to assist the electrolytic action to some extent ; but it is less
easy to discover the points of novelty exhibited by his later
apparatus. A carbon crucible, B (Fig. 77), is set in masonry. A,
and is connected by the
metal plate, 6, to a, which
is joined up to the electric
lead. A carbon rod, C,
forms the other pole, and
is joined by the holder, c,
to a screw-threaded spindle,
<7, by which vertical move-
ments may be imparted to
it with the aid of the hand-
wheel, h. The material to
be reduced is mixed with
so much carbon that it can-
not actually fuse; yet the
inventor expects the result-
ing metal to liquate from
this excess of carbon, so
that it may from time to
time be withdrawn through
the tap-hole, d. From later
publications it appears that
Willson obtained a metallic
carbide, and not metal, in
the process as he used it.
The Electrolysis of
fused Aluminium Com-
pounds by Currents of
very high Density for
Fig. 77.— The Willson furnace.
the Production of a Melting Temi>erature by the Heatix^
Action of the Current. — It is perfectly well known that, in
accordance with the laws discovered by Joule, electrical energy
in passing through any conductor is converted entirely, or in
part, into heat. The low conductance of the liquid conductors
known as electrolytes, in contrast with that of simple con-
ductors, necessitates the use of a very strong current for the
former, especially if the electrodes be not immersed to any
great depth ; and this current is indeed so powerful, that even
refractory substances may be fused and brought into the liquid
* [English Patent 9,361, June 17, 1890.]
Digitized by LjOOQIC
ALUMINIUM. 127
condition by the heat generated. Thas the whole mass of the
electrolyte becomes very uniformly heated daring electrolysis ;
and this is a great advantage as compared with the localisation
of the heat when the arc is employed.
The Hiroolt FrocesB. — The introduction of this principle
not only rendered the extraction of aluminium possible as a
manufacturing process, but solved the difficulty, hitherto in-
superable, of finding a suitable material for the construction
of the melting vessels. At the present moment all aluminium
works make use of baths, which are kept fluid by electrical
agency ; but the first and almost the only inventor who, so far, *
has given in his patent a good specification of this method of
working, accompanied by clear illustrations, is the French
engineer, Paul H^roult. In his [German] specification''^ he
describes his invention as a '* process for the production of
aluminium alloys by the heating and electrolytic action of an
electric current on the oxide of aluminium, AI2O0, and the metal
with which the aluminium shall be alloyed." [The title of the
English patent, which was taken out in the name of Henderson,
is somewhat different, and covers a larger field ; it runs thus —
** An improved Process for the Preparation of Aluminium, Alu-
minium Bronze, and Alloys of Aluminium by Electrolysis.'']
The method of heating will be sufficiently explained by the
following description and sketch of an apparatus which was
actually brought into use by the Swiss Metallurgical Company,
now the Aluminium-Industrie-Aktiengesellschaft, at Neuhausen,
in Switzerland. Fig. 78 shows a case of iron or other metal, a,
insulated below and open at the top, and provided with a thick
lining of carbon plates. A, which are held together by some
carbonaceous cementing-medium, such as tar, treacle, or glucose*
The material of which the case, a, is made must also be a good
conductor; and in order to ensure the most intimate contact
between the outer surfaces of the carbon lining. A, and the
inner side of the case, a, and hence to obtain a high conductance,
the case may, with advantage, be made by casting it as a shell
around the lining; the cooling of the mass will then produce
the desired effect. Attached to a are a number of copper pins,
a', which conduct the current, with the least possible electrical
resistance, from the negative leads to the basin or crucible, A.
In the crucible are immersed' the ends of the positive electrode, B,
the carbon rods composing which are either laid close together
or separated by small spaces, that must be filled up with a
conductor, such as copper or carbon. At the upper ends, the
* The French, English, and American Patents, some of which were taken
out in the name of his collaborators (Schweiz. Metallurg. Gesselsoh., Drey-
foae) or of his patent agent, Henderson, were— French Patents 176,711,
Apnl 23, 1886, and 170,003. April 15, 1887; English Patent 7,426, May 21,
1887; German Patent 47,165, Dec. 8, 1887 ; U.S.A. Patent 387,876, Aug.
14,1888.
Digitized by VjOOQ IC
128
ELECTRIC SMELTING AND REFINING.
carbon plates are held together by the frame, g^ which is sus-
pended from a chain by means of the eye, e^ so that the whole
electrode, B, may be brought into place, or raised or lowered at
will. The lower frame, A, attached to the electrode, B, is pro-
vided with the necessary clamps or screws for making connection
with the positive lead from the dynamo. With the exception
Figs. 78 and 79. — The Hdroult aluminium furnace in longitudinal section,
and in plan.
of the space, t, around the electrode, B, and necessary for its
free vertical movement, the melting chamber is closed entirely
by the graphite plate, k, which, however, has openings, n, for
the introduction of materials to the furnace. Channels are cut
in the lining. A, to correspond with the apertures, n, and these
channels, m, n, provide also for the escape of gases generated in
the furnace. The movable plates, o, provided with the rim, o',
and the handle, o", serve to cover the openings, n, during the
Digitized by V^jOO^ It^
ALUMINIUM. 129
progress of the melting process. The space between the graphite
plate, kj and the rim of the case, a, is filled in with charcoal
powder, as shown at k\
In starting the operation, copper, preferably in a finely-divided
condition, is placed within the crucible, A ; the sheaf of anode
carbons, B, is lowered until it comes in contact with the copper,
the current then trayerses the latter and causes it to fuse. As
soon as the bath of fluid copper, which forms the negative pole,
is ready, alumina is introduced into the crucible, and the anode
sheaf, B, is raised a little higher. The current now passes
through the alumina, which melts and becomes decomposed, the
oxygen uniting with the carbon, 6, and burning it to carbonic
oxide, which escapes as a gas from the mouth of the crucible,
and the aluminium separated from its combination with oxygen
being absorbed by the copper and forming aluminium-bronze.
The crucible is now supplied with both copper and alumina,
either continuously or intermittently, at a rate depending upon
the progress of the electrolysis.
The anode sheaf, as already explained, must be raised or
lowered according to the resistance in the circuit. This regula-
tion may be made automatic by connecting the chain carrying
the anodes, B, with a reversible electric motor, which is regulated
by the ampere-meter and acts as an electric regulator. To tap
off the melted aluminium-bronze which accumulates in the
crucible, the mould, t, lined with carbon, is wheeled into
position under the tap-hole, G; the carbon rod, c, is then
withdrawn from the latter, and is only replaced when the
mould is full. The electrolytic process is continued by lowering
the carbon, B, deeper into the crucible again, and continuing
the additions of copper and alumina as before. A current of
about 13,000 amperes and 12 to 15 volts is to be recommended
as suitable for this process.
So far the description has referred to the German patent
specification. Of the others, the English and the first of the
two French patents contain in addition the description of an
apparatus in which provision is made for external heating.
The apparatus'*^ specified in the American patent may here
be referred to (Fig. 80) since it is well adapted for use in
smaller experiments, and also possesses* some historic interest.
The crucible, a, stands on a plate, p, of conducting material.
The space between the crucible and the masonry, m, is filled
with powdered carbon, g. The carbon anode, d, passes through
the cover, 6, and the crucible itself forms the cathode.
According to the claims in the [German] patent specification
* [The current values recommended in the English patent for a fumaoe
of this tyipe, with a orocible 8 ins. (20 cm.) deep, internally, and 5) ins.
(14 cm.) wide at the top, and with an electrode 2 ins. (5 cm.) in diameter,
are 400 amperes x 20 to 25 volts.— Translatob.]
y
Digitized by
Google
130
ELECTRIC SMELTING AMD REFINING.
this process is intended for the production of aluminium alloys,
especially aluminium-bronze.
From a memoir dedicated by Regelsberger''^ to Kiliani, the
first director of the Neuhausen Aluminium Works, who
was prematurely removed by death from the field of work
in which he had laboured so successfully, it may be ascer-
tained that the Aluminium' Industrie- AktiengeseUsckcift, which
was the successor of the Schweizeritcher MetaUurgischen
GeselUchaft^ commenced, soon after its foundation in 1888, to
manufacture pure aluminium by a process worked out by
Fig. 80. — H^roult's aluminium furnace for small installations.
Kiliani in place of H^roult's aluminium alloy process, which
had previously been used there. It is much to be regretted
that, with the exception of Regelsberger's mention of the
fact that Kiliani's process agreed in principle with that of
the patents already owned by the Schweizerischer Metal-
lurgischen Gesellschaft, nothing further has been made known
concerning Kiliani's work in this direction, which without doubt
has done excellent service in the development of the aluminium
industry. Publications relating to aluminium commonly state
that the metal is extracted in America under the Hall patents,
^ZeiUchr, fur Elektrochemie^ 1895, voL i., p. 528.
Digitized by VjOOQ IC
ALUMINIUM. 131
and in Europe under those of H^roult and Minet. This state-
ment is, howeyer, at direct variance with Borchers' criticisms
of the Hall and Minet processes, as made in the first German
edition of this book, in which the American patents of C. M.
Hall and the Minet processes were both described as unworkable
under industrial conditions.
If Hall or Minet have worked out processes which differ from
those described in their patents, the author does not wish to
dispute the statement that certain factories may have one or
other of these systems in operation. The present method of
working adopted by the works in question cannot be said to
differ in any way in the fundamentally important points from
that of the H6roult process j but at that time no published
accounts, other than those in the patent specifications of Hall
and Minet, were known, and Borchers obviously could only
found his opinion on these. Nevertheless, he still adheres to
his previous opinion that the processes are not industrially
workable. In the first place, the inventors may be allowed
to speak for themselves in the actual words of their patent
specifications, which are here given, together with reproductions
of the original drawings.
The first application of Oh. M. Hall* was received in the
Patent Office of the United States of America on July 9, 1886 ;
it was divided, and formed the basis of the two patents 400,766
and 400,664 of April 2, 1889 (the day that the patent was
granted).
Hall's American Patent, No. 400,766. — '<The invention
described herein relates to the reduction of aluminium from
its oxide by dissolving such oxide in a bath containing a fused
fluoride salt of aluminium, and then reducing the aluminium
by passing an electric current through the bath, aiibstantially as
hereinafter more/ully described and claimed. In the accompany-
ing drawings. Fig. 81 represents a sectional elevation of a form
of apparatus applicable in the practice of my invention, and
Fig. 82 is a view partly in elevation and partly in section of a
modified form of apparatus.
" In the practice of my invention I prepare a bath for the
solution of the aluminium by fusing together in a suitable
crucible, A, the fluoride of aluminium and the fluoride of a
metal more electro-positive than aluminium — as, for example, the
fluoride of sodium, potassium, &c. — ^these salts being preferably
mingled together in the proportions of 84 parts of sodium
fluoride and 169 parts of aluminium fluoride, represented by the
formula, Na^Al^Fg. A convenient method of forming the bath
consists in adddng to the mineral cryolite f|f of its weight of
aluminium fluoride. The object of thus adding aluminium
fluoride is to secure in the bath the proper relative proportions
* [Compare EngUflh Patent 5,669, April 2, 1889.]
Digitized by LjOOQ IC
132
ELECTRIC SUELTING AND REFINING.
of the fluorides of aluminium and sodium. To the fused bath is
added alumina, or t,he oxide of aluminium, in sufficient quanti-
ties, and the alumina being dissoived by the fused bath an
electric current is passed through the solution by means of
suitable electrodes, C and D, connected with a dynamo-electric
machine or other suitable source of electricity, and immersed
in the solution. By the action of the electric current, which
preferably has an electromotive force of about 4 to 6 volts,,
oxygen is released at the positive electrode, C, and aluminium is
released at the negative electrode, D, which, on account of the
affinity of aluminium for other metals, is formed of carbon (/)
when it is desired to produce pure aluminium. The positive
electrode may be formed of carbon, copper, platinum, or other
Fig. 81.— The Hall furnace, U.S.A. Patent, No. 400,766.
suitable material. When formed of carbon the electrode, C, is
gradually consumed, and must therefore be renewed from time
to time ; but when formed of copper an oxide coating is formed
over the surface of the electrode. This coating serves to protect
the electrode from further destruction by the action of the
oxygen, but does not interfere materially with the conducting
qualities of the electrode.
*^ On account of the affinity of the aluminium (or other metals,
and also the corrosive action of the materials, I prefer to form
Digitized by VotOOQIC
ALUMINIUM.
133
the crucible, or melting pot, A, of metal — as iron or steel — and
protect the same from the action of the alaminium by a carbon
lining, A'. This cnunble is placed in a suitable furnace, B, and
subjected to a sufficient heat to fuse the materials placed therein,
such materials fusing at approximately the same temperature as
common salt.
"In lieu of the electrode, D (Fig. 81), the carbon lining. A',
may be employed as the negative electrode, as shown in Fig. 82,
the conductor from the negative pole of the electric generator
being suitably ( !) connected, as shown at N^, to such lining.
" In order to render the bath or solvent more fusible, fluoride
of lithium may be substituted for a portion of the fluoride of
sodium ; as, for example, for one-fourth the fluoride of sodium,
an equivalent amount of lithium fluoride by molecular weights
may be substituted. Thus 26 parts of lithium fluoride dis-
placing 42 parts of sodium fluoride, the
resulting combination contains 26 parts
of lithium fluoride for every 126 parts
of sodium fluoride and 338 parts of
aluminium fluoride.
" While I consider the proportions of
fluorides of sodium and aluminium, and
of the fluorides of sodium, lithium, and
aluminium hereinbefore stated, are best
adapted for the purpose, such propor-
tions may be varied within certain
limits without materially affecting the
operation or function of the bath, as,
in fact, any proportions which may be
found suitable may be employed. The
aluminium, as it is reduced at the
negative electrode, is melted and collects
thereon in globules, and then drops
down to the bottom of the bath, which is of lower specific
gravity than the molten aluminium, and can be removed by
suitable means ; or the bath may be poured out, and after being
cooled the aluminium can be picked out."
Clearly, in this patent, external firing is expressly prescribed
for the heating of the crucible and its contents, and carbon
cathodes are recommended for this method of work. These two
claims make it quite evident that the inventor at this time was
still far from recognising the conditions that must be fulfilled in
any process which is to be of practical use in the extraction of
aluminium.
Hall's American Patent, No. 400,664. — In this specification,
which agrees throughout with the preceding, both in idea and
expression, another solvent for the alumina is first proposed i-r-
" This combination, which may be termed the * double fluoride
Digitized by V^OO^ It!
Fig. 82. — Modified appar-
atus, Hall Patent, No.
1/ t. -r./^
134
ELECTRIC SMELTINQ AMD REFINING.
of aluminium and potassium/ is preferably formed by mixing
together 169 parts of aluminium fluoride and 116 parts of potas-
sium fluoride, such proportions of the ingredients corresponding
to the formula K^Al^Fg. A variation of these proportions
within certain limits produces only immaterial changes in oper-
ativeness of my process — as, for example (as I now believe the
fact to be), a larger proportion of potassium fluoride increases the
capacity of the bath for dissolving alumina, but at the same time
lessens its fusibility, whereas a larger proportion of aluminium
fluoride renders the bath more fusible, but decreases its capacity
for dissolving alumina.
A3
Fig. 83. Fig. 84. Fig. 85.
The Hall fumaoe, American Patent, No. 400,664.
** The combination of the fluoiides of aluminium and potassium
may be rendered more fusible and its capacity for dissolving
alumina may be increased by the addition of lithium fluoride,
or a partial substitution thereof for potassium fluoride—as, for
example, the combinations of the fluorides of aluminium, potas-
sium, and lithium, represented by the formulae KLiAlgFg or
KgLigAl^Fjg, are effective as regards their capability for dissolving
alumina, and are quite fusible."
The apparatus, shown in the annexed Figs. 83, 84, 85, to be
used in conducting the electrolysis is then described : — " The
crucible. A, containing the above-described bath or solvent, is
placed in a suitable furnace^ B, and heated sufficiently to melt
Digitized by V^jOOQ IC
ALUMINIUM.
135
the bath — i.e., to approximately a low red heat. The electrodes,
0 and D, having suitable connections with a dynamo-electric
machine, or other suitable source of electric energy, are then
inserted into the bath, and a sufficient amount of alumina placed
therein. The alumiua is dissolved by the bath, and by the
action of the electric current aluminium is reduced at the
negative electrode, D, and, being melted, sinks down to the
bottom ol the crucible."
There is a notable departure in this case from the practice
described in the previous specification, inasmuch as any material
eoniainvng carbon is here expressly excluded from use in preparing
the anodes, and cop[>er or platinum is recommended as a substi-
tute. As in the last patent, the inventor holds it possible to
use the carbon lining,
A', of the crucible (Figs.
83, 84, 85) as cathode.
In addition to iron Or
steel, copper is men-
tioned as a material
suitable for the cruci-
ble. An electromotive
force of 3 to 4 volts
still suffices for the
current.
Hall's American
Patent, No. 400,665.—
In this speci6 cation the
use of the previously
patented solvents or
electrolytes is dis-
claimed as being un-
suitable. A new recipe
is given instead, and it
is accompanied by the
description of a new
apparatus. The pre-
viously patented bath, formed of the donble fluoride of alu-
minium and an alkaline metal, '^becomes less efficient after
being subjected to electrolytic action for some time. This
change does not result from any removal of aluminium fluoride
or of fluorine from the bath, as no fluorine is separated ; but a
black or dark substance is formed in the bath, apparently from
the alkaline constituent of the same (!), which interferes with
a free electrolytic action, and increases the electrical resistance ;
hence, it has been found necessary, when employing the salts
named, to change the bath after a continuous use of the same.''
This difficulty is circumvented in the following way : — " A
double fluoride of aluminium and calcium is used, having a com-
— The Hall furnace, American
Patent, No. 400,665.
Digitized by V^jOOQ IC
136 ELKCTRIC SMELTINO AND REFINING.
position represented by the formula CaAlgFg. This composition
is formed by 169 parts of aluminium fluoride to 78 parts of
calcium fluoride or fluor spar, the aluminium fluoride being
artificially prepared by saturating hydrated alumina with
hydrofluoric acid. The fluorides of calcium and aluminium
unite in forming a double fluoride more fusible than the
fluoride of calcium and, as I beiieve, more fusible tlian either
constituent alone. While I prefer to use the double fluoride
of calcium represented by the formula CaAljF^ in carrying
out my process, there are a large number of similar double
fluorides of the same metals which may be used, and which
closely resemble the one named, and answer almost, if not
quite, as well as a bath for dissolving the alumina. Thus the
double salt of the formula CagAljFig has been successfully
employed. . . . The fluoiides above mentioned are placed
in the carbon-lined crucible or vessel, 1, arranged in the
furnace, 2, and subjected to sufficient heat to fuse the material.
Electrodes, 3 and 4, formed of carbon, when pure aluminium is
to be deposited, are connected to a suitable electric generator,
as a dynamo -electric machine. If an alloy be desired, the
negative electrode, 4, is formed of the metal with which it
is desired to alloy the aluminium. Alumina in the form of
bauxite (!), anhydrous oxide of aluminium, or any other suitable
form of alumina, ))referably the pure anhydrous oxide, A1203,
artificially prepared." . . . (The decomposition of the oxide
is described as in other specifications.) . . . "The solution
of alumina in the fused bath of the double fluoride of aluminium
and calcium is apparently heavier than the metal aluminium, and
hence this metal, if unalloyed, rises after being reduced and floats
to the surface of the bath, where it is liable to loss by oxidation
in contact with the air or with the positive electrode, where it
is subjected to a strong oxidising action ; hence, when reducing
aluminium the crucible, 1, is provided with a cover, 5, provided
with openings for insertion of the electric conductors and the
escape of gas from the crucible, and with a partition, 6, of
carbon extending down into the bath and from side to side of
the crucible, and separating the two electrodes.''
Neglecting the electrolyte for a moment, it will be seen that
this devious path has after all only led back to Bunsen's original
apparatus.
Then, again, ''As the operation continues, the cover is removed
from time to time and the aluminium removed from the surface
of the bath. . . . The specific gravity of the bath may be
lowered by the addition to it of other salts lighter than the
double fluoride of calcium and aluminium, so that the pure
aluminium will sink in the combination. . . . Thus, by the
addition to the bath above described of about two-thirds its
weight of the double fluoride of potassium and aluminium
Digitized by V^jOO^ It!
ALUMINIUM,
137
(K^Al^Fg), which is comparatively a very light salt, is formed
a combination lighter than aluminium/' Finally, in place of the
calcium-aluminium fluoride, the double fluoride of the other alka-
line-earth metals may be used. These sales of the alkaline-earth
metals are preferable to the alkaline double fluorides, because the
operation of the current is in no way affected by their saturation
with alumina.
Hall's American Patent, No. 400,666. — Here, again, a new bath
is em])loyed. It is true that good results had been obtained with
the electrolytes previously named, but the black precipitate (see
preceding specification) is undesirable. Hence, ''An electrolyte
or bath is formed of the fluorides of calcium, sodium, and alu-
minium, the fluorides of calcium and sodium being obtained in
the form of fluor spar
and cryolite respeo- ^ ^
tively, and the fluoride
of aluminium being ob-
tained by saturating
hydrated alumina
{Alo(OH)e) with hydro-
fluoric acid. The com-
pound resulting from
the mixture of the
above-mentioned fluor-
ides, which is repre-
sented approximately
by the formula
Na2Al2Fg + CaAl2r8 is
placed m a suitable
vessel, 1 , preferably
formed of metal and
lined with pure carbon,
for the purpose of pre-
venting the admixture
of any foreign material
with the bath or with
the aluminium when reduced, l^he vessel^ 1, is placed in a
furncux, 2, and subjected to suflScient heat to fuse the materials
placed therein. Two electrodes, 3 and 4, of any suitable
material, preferably carbon, when pure aluminium is desired
. . . are placed in the fused bath, or, if desired, the carbon-
lined vessel may be employed as the negative electrode, as
represented in dotted lines. . . . The reduced aluminium
sinks down to the bottom of ihe vessel, the bath being of a less
specific gravity than the aluminium."
In other details this specification varies but little from those
previously quoted, excepting that from 3 to 4 per cent, of calcium
chloride is to be added to the bath. The electromotive force is
now given as 6 volts. ,
Digitized by V^jOOQ IC
Fig. 87.— The Hall furnace, American
Patent, No. 400,666.
138
KLBCTRIC SMELTING AND REFINING.
Hall's American Patent, No. 400,667. — Ihis final patent
protects the following proposal : — A suitable bath consists of
''fluoride of calcium, 234 parts; cryolite, the xlouble fluoride
(Na^Al^F^o)* ^^^ parts; the fluoride of aluminium, 845 parts,
by weight, and about 3 to 4 per cent, (more or less) of a suitable
chloride — 0.^., calcium chloride. Alumina is then added to this
bath, preferably in sufficient quantities to form a saturated
solution. . . . Electrodes . . . are then inserted in
the bath, the negative electrode being formed of carbon when
pure aluminium is desired. The positive electrode may be
formed of carbon or other suitable (!) material."
There are here, then, repeated the possibilities and impossi-
bilities which have been described and patented previously, so
Pig. 88. — Aluminium furnace used by the Pittsburg Retluction Company
(Richards).
that there is nothing new to be extracted from these specifica-
tions unless it be the somewhat altered composition of the bath
and of the electromotive force required, which varies from 4
to 8 volts. These were the patents to which the criticisms in
the German, 1891, edition of this book were applied, criticisms
to which the author still adheres.
The author leaves it to the impartial critic to decide for him-
self whether the apparatus shown in Fig. 88 (which is, according
to Richards, that used by the owners of the Hall patents —
namely, the Pittsburg Reduction Company) has more in common
Digitized by VjOO^ li:^
ALUMINIUM. 139
with the apparatus described in the above patents or with that
of H^roult.
A later patent (1893) granted to J. B. Hall will be referred
to subseqnently.
[Practical Installation of the Hall Process.— The Pittsburg
Reduction Oonipany have put down a very large installation of
the Hall plant at their Niagara works, where it is now in full
operation. The following description, based on a paper read
by Hunt'*'' before the Institution of Civil Engineers, seems to
indicate, however, that the details of the process have been to
some extent modified since the publication of the specifications
quoted above.
The electrolytic tanks consist of iron troughs lined with
carbon, and are connected up in series. A stout copper bar
is riveted to the outride of each trough, and serves to make
electrical connection either with the anodes of the adjoining
trough or with the negative conductor of the generator,
according as the trough occupies an intermediate or (the
negative) end position in the series. The tank itself, there-
fore, with any aluminium that it may contain, acts as the
cathode. The anodes are carbon rods suspended from a copper
bar, which is placed above the vat, and are partly immersed in
the fused electrolyte ; as these carbons are gradually consumed
by the oxygen liberated in contact with them, they must be
lowered from time to time into the bath. It is found in this
inst.al]ation that the weight of carbon thus burned is approxi-
mately equal to that of aluminium reduced. The carbon linings
are not appreciably aflTected, and should (with the iron containing-
pote) last for several months in continuous use.
'Ihe process depends on the electrolytic decomposition of
alumina dissolved in the fused bath. The following rules for
guidance in the selection of a solvent have been laid down by
Hunt, and although they will for the most part be self-evident
to those who have carefully read the present chapter, they may
be quoted here : —
1. The solvent, with its charge of dissolved ore, must be fluid
at a comparatively low temperature, in order to obtain the proper
conductivity and to allow the reduced aluminium to sink in the
bath.
2. The solvent must be able to dissolve, and to retain in solution,
at least 20 per cent, of alumina at the working temperature.
3. The (thermo-chemical) heat of formation of the solvent must
be such that the latter will not be more readily decomposed than
the dissolved ore.f
♦ Trawi, In9t. Civ Eiuj., 1890, vol. cxxiv., p. 208.
t [The electrical piessiire theoretically required for the dissociation of
alumina is 2*8 volts, whilst that for the fluorides of sodium, calcium, or
aluminium is 4 volts or more].
Digitized by VjOOQ IC
140 ELECTRIC SMELTING AND REFINING.
4. The specific gravity of the solvent, when molten, must be
less than that of fused aluminium, in order to prevent shoi*t-
circuiting by metal floatiug to the top aad forming a bridge
between the electrodes, and to avoid the re-oxidation of alu-
minium at the surface.
5. There must be no solid (insoluble and infusible) bye-product
to clog the pot and to prevent (electrical) continuity.
6. The molten solvent with its dissolved ore must be a good
conductor of electricity.
7. The solvent must have no corroding action on nascent
aluminium (t.e., on the metal at the moment of deposition).
The composition of the bath may evidently be varied very
considerably without sacrificing any of these conditions; but the
mixture most commonly employed is said to consist of 677 parts
of aluminium fluoride, 251 of sodium fluoride, and 234 of calcium
fluoride. In starting the process, the ingredients are either
fused in separate vessels and transferred to the electrolyte tanks
in the molten condition, or they are melted in the tanks them-
selves with the aid of the current. When the bath is thus
prepared, pure alumina is added, and the supply should be
constantly renewed as the metal is deposited, so that the elec-
trolyte may always contain about 20 per cent, of the dissolved
oxide. The temperature of the bath is kept below 982' C. The
operation is continuous, and the aluminium is withdrawn from
the bottom of the bath from time to time, as it accumulates^
either by a siphon or by means of a ladle, in which case care
must be taken to withdraw as little as possible of the molten
salts with the metal. It is necessary that both the alumina and
the anode carbons should be as pure as possible, because all the
foreign substances present in them will pass into the bath ; and
then, since aluminium is very electro-positive in character, they
will for the most part enter into the reduced metal. The same
care need not be taken in the preparation of the electrolyte, and
the various salts used may be such as are commonly sold in
commerce. The reason of this is that the impurities (such as
silicon, lead, zinc, and copper) present in them are more electro-
negative than aluminium, and are completely eliminated by
deposition within the first two days of use. With care, the
solvent should only require renewal at long intervals, so that
after the first two days the only impurities that can be present
in the aluminium are those derived from the carbon anode and
lining, and from the added alumina.
No satisfactory information is given as to the cost of working,
nor as to the actual current employed. It is simply stated that,
theoretically, 0-7476 lb. of aluminium is the equivalent of 1000
ampere hours, and that the amount of energy required to reduce
1 lb. of the metal from alumina is about 5 H.P. hours. This, of
course, is on the as&umption that 746 ampere hours yield 0554
Digitized by VjOO^ It!
ALUMINIUM. 141
lb. of aluminium (see table, p. 9 of previous edition), and that the
electrical pressure required for diBsociation is 2*8 volts ; so that
a current of 746 amperes x 2 8 volts (%.e,, of 746 x 2-8 watts, or
of 2-8 E.H.P.) should yield 0*554 lb. Al per hour. In practice,
the author says, this is "very nearly approached." It must,,
however, be remembered that the heating of the bath is effected
by means of the electric current, which has thus to supply
sufficient energy to raise the added alumina to the temperature
of the bath, as well as to make good the loss of heat from the
whole apparatus by conduction and radiation ; and that there is
resistance to be overcome in the circuit, which also adds to the
consumption of power. Even by allowing the utmost elasticity
to the expression quoted in inverted commas, it would seem
that considerable advance must have been made since Adden*
brooke gave the yield obtained by the Hall process as 1 lb. AI
per 22 E.H.P., and since Richards stated it to be 1 lb. per 16
E.H.P. — Translator.]
The Minet-Bernard Process. — Now, as to the Minet patent,
which was subjected to similar criticism. To prevent misunder-
standings, which might arise from an incorrect comprehension of
newspaper accounts, it may be stated that Minet's and Bernard's
processes are identical. Minet's process was first put into opera-
tion at the factory of the Brothers Bernard, and the patents are
taken out in the names of these gentlemen.
The first application for a patent* was made in England in
July, 1887, when provisional protection was sought; the com-
plete specification with drawings was presented in January,
1888, and was accepted on February 17, 1888. The latter runs
thus : — ** Among the salts of aluminium, some are very slightly
fusible, while others are volatile, and in order that an electro-
lytic action may be produced it is essential that the salt to be
electrolysed should be in a particular state of liquefaction, which
is very difficult to be obtained with aluminium salts. Thus the
chloride of aluminium is too volatile, while the fluoride of
aluminium is not fusible enough.
"It has already been proposed, as regards the first, to combine
it with the salt of another metal, and to thus form a double salt
which imparts to it a little more fixity, and allows it to be
subjected at a relatively low temperature to the action of the
current in order to extract the aluminium therefirom. But the
double salt thus formed — that is to say, the double chloride
of aluminium and of sodium — is an exceedingly unstable body
which is handled with difficulty, and it gives rise to extremely
noxious vapours, hence it cannot be used commercially for the
extraction of aluminium therefrom by electrolysis ; in all cases
great losses are occasioned.
" It has been sought to render it still more fluid by mixing it
• English Patent 10.057, July 18, 1887.
Digitized by VjOOQ IC
142 ELECTRIC SMELTING AND REFIXINO.
with small quantities of chloride of sodium, and even of cryolite
{double fluoride of aluminium and of sodium); this latter body
being then adopted solely as a flux. Now, experiment shows
that if the current be passed through a bath composed of double
chloride of aluminium and of sodium, chloride of sodium, doable
fluoride of aluminium and of sodium (cryoUte), the double chloride
is volatilised to a great extent, and the fluoride of aluminium,
which is more fixed than the former, is decomposed by the
current (?).
"A bath has, therefore, been formed, in which the fluoride of
aluminium is contained in larger proportions, either in the form
of a single salt or in the form of a double salt, the Jltwride not
being employedj according to their invention, as a flux, hut being
used as the principal electrolyte, which will be subjected to the
action of the current. Experiment shows, moreover, that the
yield in aluminium increases with the proportion of fluoride of
aluminium contained in the bath. It reaches its maximum when
the fluoride of aluminium is found in a melted state without
being mixed with any other salt. Very good results are obtained
with the following proportions : — Fluoride of aluminium » 40,
chloride of sodium = 60 parts. A double salt of sodium and of
aluminium, such as natural or artificial cryolite, may even be
used, not as a flux, but as an electrolyte in the following propor-
tions : — Cryolite = 20 to 50, chloride of sodium = 80 to 50
parts. It must be understood that there can be employed as
electrolytes in the bath mixtures or compounds of fluorides of
sodium and of aluminium in different proportions to those in
which they are found in cryolite; for instance, fluoride of
aluminium = 35, fluoride of sodium = 10, chloride of sodium (or
other flux) = 55 parts.
<' Electrodes. — Their nature may vary according to the state
in which the aluminium metal is to be deposited, the following
list comprises the most general cases. For pure aluminium,
either carbon anodes and cathodes, or carbon anodes and copper
or iron (cast iron) cathodes ; for aluminium bronze, red-copper
or carbon anodes and copper or carbon cathodes; for ferro-
aluminium, iron or carbon anodes and cathodes. In short,
metal or graphite may be employed without distinction as
electrodes.
" The crucible may be of refractory earth, of plumbago, or of
metal (iron, cast-iron, red-copper, platinum, and the like), and
in cases where an alloy is required the crucible itself may serve
as the cathode, or as the anode.
*< Besides platinum, which is unalterable but costly, the other
metals are attacked by the bath (mixture of chloride of sodium
and fluoride of sodium), and have still a further defect, that is
to say, that, tohen they are evhfected to the direct action of (ke flame,
or to too high a temperature, they allow the bath to filter
through, hence a considerable loss results.
Digitized by V^OO^ It!
ALUMINIUM.
U3
** With the arrangements hereinafter described, with reference
to the accompanying drawings, snch defects are obviated. Figs.
89 and 90 show a general view of the tanks arranged in such a
manner that the bath cannot filter through. It is sufficient for
this purpose to provide the metallic tank, a 6, outwardly with a
thin brick wall.* The bath is prevented from becoming dirty
in consequence of the metal of the tank being attacked by means
of two entirely different arrangements according as it is desired
to obtain an alloy of aluminium, or pure aluminium alone.
''First arrangement. Fig. 89. For the formation of aluminium
alloys. The tank is made of the metal which forms a component
part of such alloys. A A are the carbon anodes ; a 6 is the
metallic tank serving as a cathode. At the beginning of the
operation, and as fast as aluminium is produced by electrolysis,
Fig. 89. Fig. 90.
Bernard (Minet) aluminium furnace.
there is formed on the inner surface of the tank a layer, a, of
the aluminium alloy to be obtained. When this alloy is suffi-
ciently rich in aluminium, the metal formed anew electrolyti-
cally runs to the bottom of the tank. As shown in Fig. 89, the
bottom of the said tank is slightly inclined, and is provided with
a tap-hole, T, which allows after a lapse of time, determined by
means of experiments, the withdrawal of the fused aluminium
containing smaller or greater portions of the metal forming the
tank, which is not a defect, as the object to be attained was the
formation of an alloy.
'* Second arrangement, Fig. 90. For the formation of pure
aluminium. Let A be the anode or anodes, and C be the
cathode or cathodes. The cathode, 0, is arranged in a small
* [Thifi wall is stated in the French specification as being used to protect
the vesMl/ram the coke fire employed to lyring it to a red heat, — TaiLNSLATOR.]
Digitized by VjOOQ IC
144 ELECTRIC SMBLTINO AND REFINING.
crucible, d, placed upon a plate, e, these two apparatus being
formed of fused alumina, or fluoride of calcium, or of carbon
blocks, a 6 is the metallic tank which is connected with the
cathode. The resistance, r, interposed between the cathode
and the metal tank is sufficiently great so that the derived
current in the tank represents only a small portion of the whole
current (5 to 10 per cent.). It serves to form upon the inner
surface of the tank an alloy of aluminium, and of the metal
forming the said tank, which alloy protects the latter, and is
only feebly acted upon by the bath. The metal which is de-
posited in the crucible, c?, in this arrangement is nearly pure.
** Having now particularly described and ascertained the
nature of my said invention, and in what manner the same
is to be performed, I declare that what I claim is :
" The hereinbefore described process for the extraction upon
a large scale of aluminium by electrolysis, and comprising the
extraction of pure or alloyed aluminium, the characteristic
features of said pi*ocess being :
" 1. The electrolysis of a metal bath where fluoride of alu-
minium (in the state of a single or a double salt) is used, not
as a flux, but as the main electrolyte, being subjected to the
influence of the current.
" 2. The use of a brick covering for the metallic tank so as to
avoid the filtration of the bath.
" 3. The use of the tank as a cathode for the production of an
alloy of aluminium, or at least of aluminium destined to form
an alloy with the metal constituting the tank.
*^4. The use of a derived current on the tank when it ia
required to obtain pure aluminium.
*'5. The use of fused alumina and of fluoride of calcium (or
carbon blocks) for the formation of plates, and of the crucible
serving for the production of pure aluminium."
The above specification oflers no suggestion whatever that the
bath is to be heated electrically, although this forms the very
foundation upon which the successful extraction of aluminium ia
based. On the contrary, it is much more clearly explained that
the crucible is to be protected externally by masonry against the
action of flame. The use of carbon cathodes is excluded in the
arrangement here described. The feeding of the bath during
electrolysis by the use of aluminium compounds, such as alu-
mina, was not thought of at the time that the patent was taken
out. The criticisms above alluded to apply therefore with full
weight, not only to the Hall, but to the Minet patents also.*
* [It has been stated by Ch^nevau in the Rtvtu de Physique et de Chimie
(vol. i., p. 56), that, in applying the Minet process at St. Michel, Savoy,
30 per oent. of oryolite mixed with 70 per cent, of sodium chloride is
used as a charge, bauxite and aluminium fluoride being introduced during
electrolysis. The electrical pressure required is 7*5 volts for each bath,
and the out-turn is equivalent to 0*88 oz. of aluminium per horse-power
hour. — Translator. ]
Digitized by V^OO^ It!
ALUMINIUM.
U5
It must bo specially remarked as noteworthy in the Hall
patents that an endeavour is made to find a solvent for alu-
mininm that shall be fusible at a low temperature. The fifth
claim in the Bemard-Minet patent would have been of some
weight. had not the drawing (Fig. 90) shown that the inventor
had no idea of the only right use for the material of which he
proposed to construct his crucible. A crucible made of alumina
or fluor spar after the fashion of d, in Fig. 90, would have
become dissolved by the melted bath in a very short time.
The Neuhaiisen-Froges Process. — In the year 1890 a very
carefully written, illustrated description of an apparatus that
was in use at the aluminium works then existing (Neuhausen
and Froges) went the
round of the technical
journals.* Fig. 91 shows
the arrangement in
question ; it consisted
of an iron crucible, Uned
with charcoal, and rest-
ing on an insulating
stand, through the
bottom of which was
passed a metallic nega-
tive pole. The positive
pole, made usually of a
carbon rod built up of
separate plates united to-
gether, was suspended in
the crucible from above.
A solution of alumina
in melted cryolite was
used as electrolyte, and
this was brought up to
the necessary fusion tem-
perature by the great
evolution of heat in the
bath caused by the employment of a very high current-density.
At first the metal pole inserted in the bottom of the crucible,
and afterwards the melted aluminium which was deposited
upon it, served as cathode. The crucible was charged initially
with cryolite, and later with alumina ; and then more alumina
was added as required to take the place of that which had
been decomposed by the current. The aluminium collecting at
the bottom of the crucible was run off from time to time, whilst
the oxygen of the alumina united with the carbon of the anode
to form carbonic oxide and carbonic acid.
Borohers' Aliiminlnm Extraction Apparatus. — At that
* Industries, 1890, voU viil., n, 499.
10
-Aluminium reduction. Apparatus
used at Neuhausen.
Digitized by LjOOQ IC
146
ELECTRIC SMELTING AND REFINING.
time Borchers had not succeeded in producing pure aluminium
with an apparatus constructed according to this description;
the carbon lining was always a source of disturbance, and he
therefore decided to work entirely without one, and to arrange
for the walls of the furnace to be made of the same material
as that to be melted within them. Since the source of heat
is in the midst of the furnace it was clear that, at a certain
Fig. 92. — Borchers* aluminium extraction apparatus.
distance from the zone of fusion, the cooling action of the air
surrounding the apparatus (or of a special cooling device, such
as a water jacket similar to those which are so frequently used
in metallurgical work) must leave unmelted a certain sufficient
thickness of a lining composed of alumina or cryolite. The
details of the first furnace constructed by the author on these
lines, and still built by the firm of E. Leybold's successors in
Cologne, are as follows (Fig. 92) : —
Digitized by VjOOQIC
ALUMINIUM.
U7
An iron crucible, T, with a chamotte bottom, B, is lined in-
ternally (F) with alumina or other difficultly fusible aluminium
compound. The steel plate, K, is let into the lining of the
bottom, and into it is screwed the copper tube, R, which may
be cooled by means of water or other suitable medium. A
narrow side tube, E, introduces the water for cooling, whilst the
inner tube, X, reach-
ing nearly to the top
of R, serves as an
escape pipe. The tube,
K, and, through it, the
steel plate. E, which
acts as cathode at the
beginning of the ex-
periment, are put in
connection with the
main conducting cable,
N, by means of the
clamp, V. The carbon
rod, A, forms the
anode ; it is connected
with the iron plate, U,
by means of an iron
clamp and an iron rod
which is screwed into
the plate. The current
is supplied through the
copper rod, P, and the
copper sleeve fixed on
the above-mentioned
rod. The metal accum-
ulating at the bottom
of the crucible during
the operation is run
offat intervals through
the tap-hole, S, into
the mould, G, while
the gases evolved at
the anode escape
through the openings in the cover, D, which also serve for the
admission of the charge of alumina. The crucible, T, is insu-
lated from P by a chamotte plate, I. The lining, F, in spite
of its solubility in the bath, is sufficiently protected by the
cooling action of the air playing around the walls of the
crucible. For operations of long duration, and with a high
current-density, further protection may be afforded by intro-
ducing cooling arrangements into the lining.
The above furnace can be used even with a current of about
Fig. 93.— Modified Borchers' aluminium
furnace.
Digitized by V^OO^ It!
148 ELECTRIC SMELTING AND REFINING.
50 amperes. But since, owing to its high thermal conductance,
the cooling of the metallic cathode-block cannot always be kept
within the desired limits, cooled carbon blocks, arranged as
shown in Fig. 93, were substituted in a furnace afterwards
designed for use with stronger currents.'*' This furnace had
also a sheet -copper cooling jacket, intended to prevent the
fused material reaching the metal wall of the furnace. A lining
of cryolite was stamped into place within the cooling jacket.
The use of this furnace is exceedingly simple. The anode is
first lowered until it makes contact with the cathode, and it is
then quickly drawn back sufficiently far for an arc to be struck.
A little cryolite is next thrown on to the floor of the furnace
with the aid of a spoon. Here it quickly melts in the neighbour-
hood of the arc. When the cathode has become covered with
the fused charge, the melting chamber may be rapidly filled up
with cryolite, the anode being gradually raised meanwhile, until
it is removed from the cathode by a space several centimetres
wide. Alumina is now sprinkled into the electrolyte in pro-
portion as metallic aluminium is separated out ; but at first a
slight excess of the oxide may be added with advantage, since
in this way, when the current-density is not too high, the
contact between the anode and the fused charge is improved.
It has been observed that, in electrolysing fused salts with a
high current-density, or when the furnace is too hot, the
immediate contact between the anode and the melted mass is
impaired. In this case gases are evolyed at the anode in such
quantity and at so high a pressure as to surround it with a
continuous layer of gas, through which the current can only
pass by the formation of a number of short arcs. Thus the
electrolyte becomes heated unduly, and the fault is accentuated,
especially if, as in the electrolytic separation of aluminium, the
carbon anode is constantly diminishing in size, through the
action of the oxygen evolved at its surface. This formation of
arcs, and the consequent increase in temperature, is an advantage
in the laboratory, or at lecture demonstrations, where it is un-
desirable to work with large apparatus or very considerable
quantities, because it facilitates the use of smaller apparatus
with cooling jackets around the walls. But for industrial
purposes it is obvious that this costly system is out of the
question, and that the conditions of working must be so chosen
that no heat is produced which is not required to maintain
the fluidity of the electrolyte and of the metal that is being
separated.
Bradley's Frooess. — ^The conditions to be observed in opera-
tions on a large scale will be reverted to hereafter. But first
* See Zeitschr. fur ElehtrocTiemie, 1898, vol. iv., p. 523. This furnace is
supplied by the Deutschen Gold- and Silberscheicteanstalt vorm. Bossier,
at Frsnkfort-on-the-Maine.
Digitized by LjOOQ IC
ALUMINIUM. 149
it will be well to examine certain interesting patents which
were applied for by Bradley,* in America, in 1883, but which
were granted to him only at the end of 1891 and the beginning
of 1892, and were first published as a patented process in 1896.
The most interesting portion of these descriptions is as
follows : —
"Hitherto (i.e., before 1883) this processf has been carried
on by subjecting the fused ore to the action of the current in
a crucible, or other refractory vessel, placed in a heating furnace
where the temperature is sufficiently high to keep the ore in a
melted condition ; but the greatest difficulty is encountered in
preventing the destruction of the crucible with this mode of
working the process, for it has been found that, in the case of
cryoUte especially, which is a double fluoride of aluminium and
sodium, the fused ore unites or fluxes with the crucible itself,
and that the gas liberated in the process of reduction (fluorine
gas) attacks the material of which the crucible is composed, and
the consequence is that the crucible is quickly destroyed. This
destructive fluxing action takes place to a greater or less extent
in treating almost any material, and is greatly aggravated by
the fact that the crucible is subjected to heat from without;
but even in the case of materials which do not exert a fluxing
action, the mere mechanical action of the external heat is
sufficient to make it almost impossible to prevent the cracking
of the crucible.
" The main object of my invention, therefore, is to dispense
with the external application of heat to the ore in order to
keep it fused. In order to accomplish this object, I employ an
•electric current of greater strength or intensity than what
would be required to produce the electrolytic decomposition
alone, and I maintain the ore or other substance in a state of
fusion by the heat developed by the passage of the current
through the melted mass, so that, by my invention, the electric
current is employed to perform two distinct functions, one of
these being to keep the ore melted by having a portion of its
electrical energy converted into heat by the electrical resistance
oflered by the fused ore, and the other being to eflect the desired
electrolytic decomposition, by which means the heat, being pro-
duced in the ore itself, is concentrated at exactly the point
where it is required to keep the ore in a state of fusion.
" Another feature of my invention consists in dispensing with
the crucible for holding the ore, and in employing a body or
heap of the ore itself to constitute the vessel or cell in which
the reduction takes place, which is not destroyed by the chemical
action of the fused ore and the gas liberated, and which, there-
* U.S. A. Patents 464,933, granted Dec. 8, 1891 ; 468,148, Feb. 2, 1892;
and 473,866, April 12, 1892.
t [This quotation is from U.S. Patent 468,148.— Tbanslatob.]
Digitized by V^OOQ IC
150
ELECTRIC SMELTING AND REFINING.
fore, admits of the process being perfectly continuous, nothing
being required but the charging of fresh ore as fast as the
reduction goes on, either from without or from the sides or
walls of the heap itself.
** To enable others to carry out my process, I will proceed to
describe it as applied in one particular case to the extraction of
aluminium from its ore cryolite.
" Upon a hearth of brick * or other suitable material is piled
a heap or body of the ore more or less pulverised, in the shape
of a truncated cone, and a cavity or basin is excavated in the top
of the heap to contain the fused portion of the ore which is to be
treated electrolytically. In order to fuse the ore at the start, I
take two electrodes of a suitable material, such as already used
in like processes where fusion has been effected by an external
furnace, and connected, respectively, to the two poles of a
dynamo electric machine or other source of current, bring the
said electrodes into contact, separate them sufficiently to produce
Fig 94.
Fig. 95.
Bradley's aluminium smelting furnace.
an electric arc, and then thrust them into the ore lying at the
bottom of the cavity or basin, where the ore soon fuses by the
heat of the arc and becomes a conducting electrolyte, through
which the current from the electrodes continues to flow. The
arc, of course, ceases to exist as soon as there is a conducting
liquid — the fused ore — between the electrodes, and the passage
of the current then takes place through the fused ore by conduc-
tion, and the heat is produced as it is in an incandescent lamp.
The arc is merely used to melt the ore in the beginning, and the
ore is kept melted by incandescence, so to speak, the metallic
aluminium being gradually deposited at the cathode and the
fluorine gas being set free at the anode so long as the ore is
maintained in a state of fusion. As soon as the action is properly
started the electrodes should be moved a little farther apart, in
order that the metal set free at the cathode shall not form a
short circuit between the electrodes or be attacked by the
fluorine set free at the anode."
* In this paragraph reference letters, here omitted, are given in th&
original specification to two figures not reproduced in this volume.
Digitized by
Google
ALUMINIUM.
151
Figs. 94 and 95 show two forms of the smelting arrangement.
In Fig. 94, 2 is the heap of ore piled on a carbon plate, 7 ; 5 is a
source of current connected with 7 and with the electrodes, 4
and 4' ; 3 is the electrolyte, from which the electrode, 4', may be
removed after the process has fairly started, as it is only required
to produce the arc initially required to begin the fusing process.
Fig. 95 shows the same furnace without the electrode, 4', and its
conductor. The arc necessary at first is, in this case, struck
between 4 and 7.
Bradley's process and apparatus undoubtedly surpass all
others in simplicity. In principle
they agree with Borchers' con-
struction of furnace and method
of working.
Kiliani's Fumaoe. — A patent
specification of Kiliani's,* who at
the time at which it was granted
was Director of the Aluminium^
Industrie - Akliengesellschajt, de-
scribes an aluminium furnace, but
says nothing of the lining of the
melting hearth, or of the arrange-
ment of the cathode.
The columns, a (Fig. 96), sup-
port the vessel, 6, which is in-
tended for the fusion of the elec-
trolyte; the cathode enters this
vessel below, and makes connec-
tion with the negative pole of the
generator by the conductor, c. The
cross piece, d, carries the arrange-
ment for giving support and motion
to the positive electrode, e. Within
the sleeve, c^, is the anode spindle,
/, which in vertical section has the
form of a rack, so that it may
engage with the pinion, g, the
latter being actuated by the worm
and wheel, h h^, in connection
with the handwheel, h^. The teeth and notches in the spindle,
/f take the form of rings and grooves respectively, which are
cut round the whole of its circumference, so that if turned
upon its axis it may in any position act as a rack, and thus be
still in gear with the pinion, g. A worm wheel, k, is so keyed
on to the spindle, /, that the latter is free to move in the
direction of the bore of the wheel. The wheel, k, gears into
Fig. 96. — Kiliaiii's furnace.
• German Patent 60,508, April 21, 1889.
20, 1889.]
[English Patent 6,745, April
Digitized by VjOO^ It!
152
Figs. 97 to 101.— Schindler's aUiraininm furn^V^OOgle
ALUMINIUM. 153
the worm, t, which is mounted on the axis of a driving pulley.
Electrical connection with the anode is effected by means of
branches in the boring of the spindle,/. When this mechanism
is in action, the anode will be maintained in constant rotation
by means of the worm wheel, while vertical motion may be
imparted to it in either direction by the hand-wheel, h\
Schindler'B Furnace. — Schindler, who succeeded Kiliani,
patented, in 1896, a furnace* with electrodes that could be
cooled. In the accompanying figures (Figs. 97 to 101), A refers
to the carbon anode ; B to the anode-holder (which may be con-
structed in one piece, or in several parts, h) ; B^ is the holder for
the cathode, A^, in the case when the latter is not connected up
to the current by a simple plate, P ; c and c^ are cooling ducts in
B and B*; r, r^, r^, r^, ?-*, and r^ are pipes for the conveyance of
the cooling medium to and from the ducts ; 0 and C^ are the
cables connected to the electrical generators ; and s s are screw
bolts for joining up the two parts of the anode-holder. The
arrangement of the cooling ducts is similar to that in the
Liirmann slag moulds.
Grabsu's Process. — Qrabau,t in 1891, patented the electro-
lysis of a fused mixture of aluminium fluoride or oxy fluoride
with alkali-metal carbonates, which was to yield alkali-metal
fluoride or cryolite as a by-product. The nature of the reaction
is shown by the following equations : —
2A1^, +
6Na,C03
4ALF, +
6Na.,C0,
2A1,0F4 -f
4NajC03
SAljOF^ +
3Na,C0a
-1-30 =
4A1
-1-
12NaF + 9C0,
{Anode)
(-)
(+)
+ 3C =
4A1
+
2(ALA.6NaF) + QCOj
iAnodf)
(-)
(+)
+ 3(; =
4A1
+
8NaF -f- 7C0j
{Anode)
(-)
(+)
+ 3C =
4AI
-f-
AL^Fe-BNaF + 6C0j
(Anod€)
(-)
(+)
It must be remarked, however, that on merely melting to-
gether a mixture of aluminium fluoride and alkaline carbonates,
an active disengagement of carbonic acid is observable, probably
in accordance with the equation : —
AljFa + 3Na,C03 = Al^O, + 6NaF + SCOj.
So that in this instance it is a solution of alumina in melted
alkaline fluoride that is really submitted to electrolysis.
The J. B. Hall Process. — In the year 1893, J. B. Hall (not
the C. M. Hall whose patents have been discussed previously)
took out a patent % in America for a process in which the old
mistake of using a carbon-alumina anode is repeated. The
crucible itself is to be used as cathode, and is to be made of the
♦ U.S. A. Patent 673,041, Dec. 16, 1896.
t German Patent 02,861, 1891. [English Patent 10,000, June 13, 1891.]
X U.S. A. Patent 603,929, Aug. 22, 1893.
Digitized by LjOOQ IC
154 ELECTRIC SMELTING AND REFINING.
same materia), which, on account of its want of durability, must
be enclosed within an iron shell. The electrolyte is a fused
mixture of aluminium, sodium, and lithium chlorides, in which
the bath is replenished at the expense of the alumina contained
in the anode. The inventor has also obtained moderaiely siUiS"
factory results (!) by omitting the alumina in the anode com-
position, and adding it direct to the bath instead. His '^new ''
method of producing alumina is also worthy of note. He treata
aluminous material with sulphuric acid, and heats the resulting
hydrated aluminium sulphate, thus breaking it up and obtaining
the oxide !
EleotrolyBis of Aluminium Sulphides. — Two patents were
taken out in Germany in 1890 for the production of aluminium
by the electrolysis of aluminium sulphide dissolved in fused
alkali-metal chlorides ; but the process is only of interest from
the theoretical standpoint. Both patents were granted on the
same day (Nov. 18, 1890). The patent which received the
earlier number was granted to Bucherer,* and is concerned
mainly with the production of the sulphide, while the other,
taken in the name of the Aluminium- Indttstrie-Aktiengetell-
achaft t sets forth the advantages accruing from the use of this
electrolyte.
Bucherer claims that, *' by the action of the sulphides or poly-
sulphides of the alkali- and alkaline-earth metals, with the
addition of sulphur and carbon in excess, aluminium oxide or
hydroxide, when heated, forms double sulphides of aluminium
and alkali- or alkaline-earth metals," the reaction being illus-
trated by the following equation : —
3Na,S + ALjOj + 3C + 3S = NagAl^SB + 3C0.
In a later account,^ he describes experiments on the con-
version of the aluminium oxide into the sulphide without the
agency of other sulphides. A mixture of alumina and carbon
was heated to a white heat in clay retorts, and was completely
converted into sulphide by means of sulphur vapour, which waa
allowed to operate for a sufficiently long period, thus : —
ALjOj, + 3C + 3S = 3C0 + A^?,.
The sulphide so obtained could then be dissolved in a fused
alkaline chlonde, and be electrolysed by a current of low electro-
motive force.
The Aluminium ' Industrie - A ktiengeselUcha/t use a similar
mixture which is brought to a state of fusion, either by means
of the electric current, in which case an E.M.F. of 5 volts is
necessary ; or by external heating, when only 2^ to 3 volts will
be required for the electrolysing current. Cast or wrought-iron
* German Patent 63,995, Nov. 18, 1890.
t German Patent 68,909 [? 63,999], Nov. -18, 1690.
t Zeitschrift fiir auyewaudte Chemie^ 1892, p. 483.
Digitized by V^jOOQ IC
ALUMINIUlf. 155
vessels may be used, and they should be lined with carbon. It
is claimed as a special advantage of this process that neither this
lining nor the carbon anodes has any prejudicial efiPect upon the
materials which are here brought into contact with them ; and
that on this account, besides effecting a considerable saving
of carbon in the various parts of the apparatus, a very pure
aluminium is obtained.
The behaviour of the sulphide electrolyte is so remarkable,
that it is a matter for regret that at present no advantage can be
taken of it, because up to the present time no sufficiently cheap
method has been invented for the production of aluminium
sulphide. It must not be forgotten, on the one hand, that the
metal may be obtained directly by electrolysis, and on the other,
that the sulphide must first be prepared from the pure oxide,
whilst as Bucherer himself points out, there is scarcely a purely
chemical reaction known which requires so great an absorption
of heat as that above described ( Al^Og + 30 + 3S = AlgSg + 300).
Finally, the author, speaking from the experience of many years
in the manufacture of alkali- and alkaline-earth metal sulphides,
points out that the production in gross of pure aluminium
sulphide, or of the double sulphides of aluminium and the
metals above named, is an exceedingly difficult problem. With
the means at present available on a working scale, the possibility
of introducing the electrolysis of sulphides for the manufacture
of aluminium is bound up with the solution of this problem.
Many processes have since been patented for the production
of pure aluminium sulphide, or for the lessening of the cost
of electrolysis by the introduction of sulphides into molten
aluminium baths (q/I Blackmore, PeniakofT, and others) ; but it
will suffice to touch very lightly upon them. In the same way,
there are other processes, which it is quite unnecessary to
describe, that have been patented since the appearance of the
last edition of this book. On the other hand, several very
valuable papers have been published relating to the management
of aluminium works, and these have given a very clear insight
into the method of working on the large scale.
Preliminary Treatment of Bauxite and Oryolite. — It has
already been stated that none of the minerals used in the
production of aluminium can be employed in their natural
condition. The raw materials are cryolite and bauxite. Of
these, the former serves chiefly as the flux or solvent for the
pure alumina which is obtained from bauxite. The cryolite
can be used direct in selected pure specimens ; or in a less pure
condition after a short electrolysis, since the greater part of the
impurities present would thus be separated out first.
Two methods are used for the preparation of bauxite. In the
older process, bauxite is calcined in admixture with soda in a
reverberatory furnace until the charge begins to sinter together.
Digitized by VotOOQIC
156 ELECTRIC SMELTING AND REFINING.
£j this means the iron present is converted into insoluble oxide
and the aluminium Jiydroxide into sodium aluminate, which
latter can then be extracted by treatment with water. The
alumina is then separated from the clear solution in the form of
pure hydroxide by the use of carbon dioxide, which also recon-
verts the soda into carbonate. The aluminium hydroxide, after
it has been filtered, washed, and dried, is converted into the
anhydrous oxide by roasting. The newer process, which was
introduced by Bayer, is employed in the alumina department of
the n^roult Works. Sutherland* has described before the
Institution of Mechanical Engineers the application of this pro-
cess in the Larne Harbour Works of the British Aluminium
Company, in Ire' and.
The bauxite used was obtained from the County of Antrim,
and had the following average composition : —
Aluminium oxide, .... 56 per cent.
Ferric oxide, 3 ,,
Silica, 1*2 ,,
Titanic acid, 3 ,,
Water, 26 „
100 ,.
Two Lancashire boilers are used, each about 26 feet long and
8 feet in diameter, working at a pressure of 7 atmospheres.
These serve an 80-H.P. steam engine, the exhaust steam from
which is applied to the heating of tbe evaporating pans. Disin-
tegrators are employed for pulverising, and the crushed material
is thrown on a sieve, with four meshes to the linear inch, which
delivers the fine particles to the calciner and returns the coarse
material to the disintegration.
The object of calcining is to destroy the organic matter con-
tained in the bauxite, and to convert the iron completely into
peroxide (Fe203) ; but the roasting is conducted at the lowest
possible temperature in order that the solubility of the alumina
may not be impaired. The calciner consists of an iron tube
33 feet long and 3J feet in diameter, lined with fire-brick, and is
provided with a fire-grate, dust-chamber, chimney, and cooling-
tube. Both the calciner and the cooling-tube are mounted on
rollers, so that they can be rotated, and are placed at an inclina-
tion of 1 : 25. The cooling-tube is about 30 feet long and 2h
feet in diameter, and is cooled while in use by a current of cold
air passing through it. The calcined material is conducted by
a spiral conveyor to a second crusher where it is pulverised
until the material has passed a sieve of 30 meshes per linear
inch. The fine material is collected in a receiver, from which it
is carried away in trucks as may be required.
• Transaetiwi Iitst. of ^Uchan^cal Engineer a^ 1896, and Bnffineering,
1896, vol. Ixii., p. 291.
Digitized by LjOOQ IC
ALUMINIUM. 157
The bauxite so prepared is now treated with a caustic soda
solution of 1*45 specific gravity at a pressure of from 5 to 7
atmospheres in iron pressure kiers of about 11 feet in length and
5 feet in diameter, with walls about { inch thick. The steam
used for heating is passed into an outer jacket, and in a short
time brings the charge (3 tons) in the inner cylinder up to a
pressure of 5 atmospheres, which is maintained for from two to
three hours, the mixture being constantly stirred throughout.
The mud is then discharged (by its own pressure) into a tank
placed at some height from the ground. It is here diluted
until its specific gravity is 1*23, and is then allowed to run
into filter-presses. Each filter-press contains 50 compartments
measuring 2 feet 6 inches x 2 feet 6 inches x 1 inch. The
filtrate is not yet sufficiently clear for use, and is therefore
passed through a filter of wood-pulp contained in lead-lined
wooden tanks, 10 feet long, 6 feet broad, 3 feet deep, with a
sieve of ^-inch megh supported inside on a ledge about 6 inchea
from the bottom. The sieve carries about 50 lbs. of wood-pulp
previously brought to the condition of a thin paste by boiling-
with water. Two sieves of this kind are superposed.
The clear liquor is now pumped into vertical and cylindrical
decomposing tanks, built of iron plates, about 20 feet high and
13 feet in diameter, provided with agitators. Instead of being
treated here with carbon dioxide, as by the old method of
working, the solution is, in Bayer's process, stirred for 36 hours
with pure aluminium hydroxide from a previous charge. In this
way about 70 per cent, of the dissolved alumina is precipitated ;
the deposit is allowed to settle, and the clear solution is run oiT
into another vessel. The hydroxide mud is pumped into filter-
presses which are filled under a final pressure of 5 atmospheres.
The press cakes are washed in the press, and the admixed water
is discharged by means of air-pressure while still in situ.
The cakes are calcined in a reverberatory furnace, of which
the hearth is 20 feet long by 6 feet wide, heated by producer-
gas. Although the water is expelled at a low temperature, the
alumina is heated to 1,100* C, since, if this were not done, the
oxide would take up water again in store, and would then be
unsuitable for electrolysis.
The dilute caustic soda liquors are concentrated in a three-
chambered vacuum evaporator, until they are again brought to a
specific gravity of 1 *45, at which strength they are again service-
able for treating fresh bauxite. The condensed water from the
evaporator, as well as the dilute wash-waters from the filter
presses, serve for the dilution of the slimes and of the solutions
of caustic soda, whilst the condensed water is also available for
feeding the boilers.
The Indnstrial EleotrolyBls of Alumina.— The electrolysis
of aluminium oxide is again, as it was in the earliest experiments,.
Digitized by V^OOQ IC
158 ELBCTBIG SMELTING AND REFINING.
conducted in iron vessels lined with carbon. The space avail-
able for the fused charge varies with the current-strength to be
used up to 2 feet 6 inches by 5 feet. With these dimensions a
current of 8,000 amperes may be used, which, reckoning only
the internal area of the bottom of the vessel, corresponds to a
current-density of 7,000 amperes per square metre [650 amperes
per square foot]. The earliest accounts of the H^roult process
described the use of a bath containing a very large proportion
of alumina and, in consequence, requiring a higher temperature
than is now used, and the current-density then used was said to
be from 23,000 to 25,000 amperes per square metre [2,150 to
2,300 amperes per square foot]. Whilst formerly it was cus-
tomary to use a bath of alumina to which cryolite was added as
a flux, a cryolite bath is now employed, to which (according to
an account given by Wallace,* of the Aluminium Works of the
British Aluminium Company at Foyers in Scotland — where the
H^roult system is used) aiumina is added in the proportion
necessary to replace that which is decomposed in the electrolytic
deposition of aJuminium. The published statements as to the
electromotive force used and current-efficiency obtained are
very discordant. The more favourable accounts give the poten-
tial difference at the terminals of the bath as 3 to 5 volts, whilst
others give it as 9 to 10 volts. Having regard to the fact that
the voltage required for the decomposition of the alumina
(calculated from the formula, E = jr-^r; t^^tk^) is
n X 0-24 X 96,537/
Approximately 2*8, it is perhaps possible that, under favourable
conditions, a difference of potential of 5 volts may be observed
between the electrodes; but it is hardly to be expected that
this could be attainable as an average reading in practical work.
On the other hand, a potential difference of 9 volts is cei*tainly
too high, with the low current-densities now used. An average
of 7-5 volts with a current-density of 7,000 amperes per metre
{650 amperes per square foot] and a current-efficiency of 90 per
cent, probably represents the results in actual work on a large
«cale. From this it may be calculated that at least 1,400 E. H. P.
must be applied in order to produce a yield of 1 ton of alu-
minium per day of twenty-four hours, a figure which agrees
very closely with that given in the last edition of this book as
having been calculated from the experimental results obtained
with a machine capable of giving only 2 E.H.P. The author
had at that time determined that about § kilogramme [1*47 lbs.]
might be expected to be deposited by 1 KH.P. in twenty-four
hours. Hence a little less than 1,500 H.P. would be necessary
to yield 1,000 kilogrammes of aluminium in tweniy-four hours,
whilst the figures since quoted as obtained from works' practice
vary between 1,400 and 1,600 E.H.P. for the same output. As
*Jounu Soc. Ckem, Ind,, 1898, p. 308.
Digitized by LjOOQ IC
ALUMINIUM. 159
above explained, therefore, it is most probable that a consump-
tion of 1,400 H.P. must be reckoned for an output of 1 metric
ton per twenty-four hours.
The progress that has been made in the manufacture of car-
bons for electrodes, and the low temperature— scarcely higher
than 750* G. — at which it is possible to maintain the bath under
the modem conditions of low current-density, make it possible
to dispense with the artificial cooling that, under other circum-
stances, was necessary to maintain the metal or carbon cathodes
and the cryolite-, alumina-, or carbon-lining of the furnace, and
to prevent the contamination of the electrolyte and of the metal
under deposition by impurities derived from the materials of
which these portions of the apparatus were made. The special
cooling appliances need not now be used, provided that there is
nothing to check the free circulation of air around the apparatus.
Indeed, with the dense carbon of high thermal conductance now
used for the crucibles, it may even be necessary, at least in the
upper portions of the side walls, to interpose between the metal
walls and the carbon lining a layer of material of low conduc-
tivity, snch as fireclay, or magnesia, or masonry.
The cryolite is melted on the large scale exactly as it is in the
small experimental apparatus. On the small scale, however,
and especially for purposes of demonstration, it is seldom pos-
sible to maintain the process long enough to obtain sufficient
metal to render it necessary to tap the furnace. On this account
the author has not referred to this part of the process in describ-
ing the use of the smaller apparatus. With such a plant, it is
usually best to let the contents of the furnace cool, and then to
break the globules of aluminium out of the solidified mass. It
is, however, self-evident that, with a large plant working on
long runs under industrial conditions, the metal collecting in
the crucible of the furnace must be run off from time to time
through a suitable tap-hole, exactly as is customary with all
metallurgical furnaces for the smelting of metals; and the
manner of conducting this part of the process is so well under-
stood that there is no need to refer to it further here. On the
other hand, it is not at first sight so readily understood why
aluminium separates at the bottom of a bath of fused cryolite,
for the numbers commonly quoted as representing the specific
gravities of the two materials would lead one to predict the
contrary.
The specific gravity usually of aluminium is given as 2*7, and
that of cryolite as 3. J. W. Richards has, however, experimen-
tally determined the densities of the materials employed both in
the molten condition and after cooling. The results, which
suffice to explain the apparent anomaly, are given in the follow-
ing table : —
Digitized by VjOOQ IC
160 ELECTRIC SMELTING AND REFINING.
SPECIFIC GRAVITIES OF ALUMINIUM COMPOUNDS.
Specific GravitiM. ,
Melted.
Solid.
Commercial aluminium
Commercial Greenland cryolite,
Cryolite saturated with alumina,
Cryolite with aluminium fluoride, ALF- . 6NaF
+ 2AlaFg = 3(Al2F6.2NaF), .
The same mixture saturated with alumina,
2-54
2-08
2-35
1-97
214
2-66 !
2-92 ■
2-90
2-96
2-98
The Conditions most fayourable for Altuniniom Ex-
traction.— On comparing the foregoing accounts of the extraction
of aluminium on both the industrial and the experimental scale^
it is possible to lay down the following conditions for the extrac-
tion of aluminium : —
1. The successful extraction of aluminium by electrolytic
means is only possible when anhydrous fused aluminium com-
pounds are used. (Determined by Bunsen in the year 1854.)
2. The proportion of aluminium in a fused bath undergoing
electrolysis may be maintained constant by the addition of
aluminium oxide during the process of electrolysis. (Ascertained
by St. Claire Deville in the year 1854.)
3. The heat required to melt che charge is produced by the
current used for electrolysis. The aluminium industry is
founded on this condition of working, which was described in
the patent specification of H6roult in 1887, and in that of
Bradley, which was published in 1892, although the patent itself
had been applied for in 1883. The temperature of the bath is
750' C.
4. The above condition (3) makes it possible to use an iron
melting vessel, lined with aluminium compounds (alumina or^
better, cryolite) or carbon, and kept cool on the outside.
6. For electrode materials there are available — For the cathode^
metal or carbon blocks so arranged that they are so far cooled,
either by air or by some other cooling medium (preferably water),
that they do not cause the addition of any impurity to the
aluminium deposited : for the cmode, carbon blocks, which, if of
large size, must be built up of smaller blocks or plates of similar
material.
6. The current-density should be 7,000 amperes per sq. metre
S650 amp. per sq. fb.] of horizontal sectional area of the bath.
From Wallace's paper, 1898.)
7. The potential difference required, allowing for drop of
potential owing to the resistance of the conductors and to con-
Digitized by V^jOOQ IC
ALUMINIUM. 161
tact resistance, is on the average 7*5 volts per furnace, or, at
times, somewhat less than this.
8. The absorption of power should be 1,400 electrical horse-
power for a yield of 1 ton [1,000 kilogrammes] per day of 24
hours. In the last edition of this work (1895) this number was
given as (probably) 1,500 H.P., as the result of experiments
made by the author.
The above conditions are based on the more trustworthy of
statements published up to the year 1898. Since then other
accounts of aluminium extraction have appeared, but there is
practically very little to be learned from them. Kingmann^
and Chandler t have published accounts of the process as applied
in the works of the Pittsburg Reduction Company, but the two
accounts are more or less at variance in some of the particulars.
Haber and Gkipert J have endeavoured, from experiments made
on a small scale, to lay down conditions for the industrial pro-
duction of aluminium, but they have assumed the use in works
of the obsolete current-density of 25,000 amperes per sq. metre.
Alumlninin Works in Operation. — Aluminium is now
produced in the following works : —
The Aluminitmi' Industrie- A ktiengeMllschaft are the oldest
manufacturers of the metal, and own three works at the present
time. The oldest of these, in Neuhausen (Switzerland), usea.
4,500 H.P.; the newer works, in Bheingelden (Baden), over
5,040 H.P. ; and those in Lend, near Q«stein (Austria), over
7,500 H.P. In each of these works, however, a part of the
power is applied to the production of carbide, sodium, and other
substances.
The Societe Electr(HneUdlurgique Frangaise possesses four
sources of water-power — viz., at Froges, near Champ (Fs^re);
near Gardaune, at the mouth of the Rhone ; and near La Praz,
in Savoy. The latter utilises nearly 13,000 H.P. in all, which
is mainly used in connection with the manufacture of aluminium,
including, however, the production of electrodes, the rolling
and further mechanical treatment of the metal, and, of course,
the lighting and various power requirements of this large and
complex works. After satisfying these requirements scarcely
8,000 H.P. remains available for electrolytic purposes, and
even this is not all applied to the production of aluminium.
Alter completion of the extension for the utilisation of the
above sources of power the company has available at least
28,000 H.P.
The Compagnie des ProduUs Chimiques cFAlaU et de la
Camargue, which had used the St. Claire-Deville sodium-process
for the reduction of aluminium from 1859 to 1889, has, since
* Western Electrician, March 17, 190a
fJaum. 8oc. Chem. Ind., 1900. vol. xix., p. 609.
t Zeitsehr.far ElekfrocJiemie, 1902, vol. viii., pp. 1 and 26.
Digitized by
Google
162
ELECTRIC SMELTING AND REFINING.
1897, preliminarily applied water-power equivalent to 3,500
H.P. to the electrolytic extraction process.
The American works of the FiUshurg Reduction Company have
latterly been completely transferred to the Niagara Falls, where
the Company is utilising 10,000 H.P. If, as is probably the
case, a part of this is employed for lighting and power, including
the farther treatment of the metal, about 6,000 H.P. remain for
use in electrolysis.
Finally, the British Aluminium Company^ which was the last
to commence the manu&cture, when employing the full power
of the plant available at the Falls of Foyers (Scotland), uses
about 5,000 H.P., which is, however, exclusively applied to
electrolytic work, since the Company possesses a separate
electrode factory at Greenock, and a rolling mill at Milton.
Uses of Altuninium. — The steadily increasing use of alumin-
ium is due to its valuable chemical and physical properties.
The fact that the strength of the metal is but slight, and rapidly
diminishes when the metal is heated even but slightly, militates
against the extension of its use as a material for construction,
concerning which, on account of its low specific gravity, great
hopes had been raised, after the difficulties connected with its
xnanufacture had been overcome. But it is not yet universally
recognised that aluminium is now one of the cheaper metals.
If, however, its specific gravity be taken as unity, and its price
be compared with the prices of equal bulks of other metals for
which aluminium could in one way or other be substituted, the
proportions given in the following table will be obtained : —
Specific
Price in
Price of equal bulks in
Metal.
Gravity.
Marks per
Shillings
Marks per Shillings
kilogramme.
per lb.
kilogramme. per lb.
Copper, .
3-37
10 -1-5
0-46.0 -68
3-37-6-06 1 1 63-2-29
Tin,. . .
2-76
2-4 -2-7
109-1 -23
6-63-7-46 3-01 3-39
Brass,
316
1-0 -1-2
0-46-0-66
316378 1-43-172
Aluminium,
1-00
2-00
0-91
200 i 0-91
Lead, . .
4-30
0-24-0-36
011-016
1-03-1 -66 , 0-47-0-71
Zinc,
2-70
0-35-0-40
017-018
0-95-1-08 0-43-0-49
Aluminium in the form of sheet wire and tube is used as a
substitute for copper, brass, bronze, zinc, and silver, in their
most varied applications. The wire has been successfully used
in many places for electrical conductors. The sheet is employed
for cooking- and table utensils, preserving-pans, articles of mili-
tary equipment, combs, brushes, bicycle components, and in
place of wood in ship- and waggon-building, book -covers, book-
shelves, coffins, and the like.
Digitized by VjOOQ IC
ALUMINIUM. 163
The very considerable shrinkage (1-7 to 1*8 per cent.) of the
metal during solidification was, at first, a hindrance to the use
of aluminium for castings. To obviate this a small percentage
of a so-called hardener was added, such as, for example, copper,
nickel, zinc, manganese, tin, chromium, titanium, tungsten, or
vanadium. Aluminium so alloyed is sometimes known in the
United States as Dickel-aluminium. In the Paris Exhibition of
1900 there were, however, cast-goods and art- wares (e.g,^ busts
of St. Olaire-Deville and Wohler) satisfactorily made of pure
aluminium. The above-mentioned difficulties are now obviated
by using large runners and rising-heads in the moulds used for
the castings. In melting aluminium for casting no flux should
be added. If this is attended to, there is no fear of the metal
taking up impurities from the carbon of the crucible, or from
the silica of the clay that is mixed with the graphite, provided
that the temperature is not allowed to rise too high above the
melting point of the metal. But if the metal be overheated,
even in the absence of fluxes, it will take up both carbon and
silicon from the walls of the crucible. Owing to the porosity of
the vessels used for melting the metal, the fuel employed for
heating them should be as free as possible from sulphur. For
this purpose wood-charcoal is best, but a pure coke containing
but little sulphur may be substituted ; and natural-gas is some-
times used in the United States. Aluminium readily dissolves
gases, such as nitrogen and the hydrocarbons, at temperatures
but little above its melting point; and nitre is added to the
metal to expel these gases shortly before pouring. This is done
by wrapping the nitre (about an egg-spoonful to 50 kilogrammes
[1 cwtwj of metal) in a piece of moistened writing-paper, throwing
it into the crucible after removing the latter from the fire, and
quickly forcing it down to the bottom of the pot so that, in
rising to the top, it must pass through the whole depth of the
metal. A brisk action is set up, and, during the period of this
reaction, all the dissolved gases are expelled. The charge is
poured at as low a temperature as possible into iron or sand
moulds, provided, as above stated, with large runners and rising-
heads.
Aluminium castings are used for table- and cooking-utensils,
sides of baths, surgical instruments, parts of electrical apparatus
(in this case usually alloyed with 5 per cent, of copper, and as a
substitute for brass), bells, such parts of machinery for use on
board ship as are not required either to be especially strong or
to be heated, and as a substitute for wood, for supports which
are not intended to carry great weight, and for sheathing and
plating in ship- and waggon-construction.
Aluminium is very largely used in the preparation of alloys ;
thus, for example, aluminium-bronzes are made with 2*5, 5, 7*5,
and 10 per cent, of this metal. The direct production of these
Digitized by V^jOOQ IC
164 ELECTRIC SMELTING AND REFINING.
alloys by the simultaneous reduction of aluminium and copper
compounds, as formerly carried on in the Cowles process, has
now been abandoned, since, at the present prices of the two
metals in question, alloys of this kind are best prepared direct
from the constituents at the place at which they are required.
To do this the copper is first melted in a plumbago crucible,
using the same precautions as are recommended aboye for the
casting of aluminium ; after the copper is melted the crucible
is seized, or lifted out of the fire, and the desired quantity of
aluminium is added. The crucible becomes momentarily cooler,
owing to the absorption of heat by the aluminium as it fuses ;
but after that there is a very considerable evolution of heat ;
and if the crucible is still in the furnace it must be quickly
removed. The charcoal is then taken off the surface of the
metal, and the charge is rapidly stirred and poured. It is best
first to run it into bars which can be re-melted as required for
casting purposes.
A very large number of alloys besides bronze have been
brought before the public during the past ten years, but for
an account of these reference must be made to works treating
specially of aluminium. Quite recently an alloy of magnesium
and aluminium termed maffnalium has been placed upon the
market, and this alloy is of special interest.
The applications of aluminium in the casting of other metals
are based, not upon the formation of aluminium alloys, but
upon the chemical action of the metal. Many metals, such as
iron, copper, and nickel, possess the property of alloying with
their lower oxides (FeO, CugO, NiO) ; hence, in smelting,
refining, or melting them, contact with the air or with other
oxygen compounds leads to the formation and solution of these
lower oxides, and, on the other hand, their complete removal is
difficult even although reducing substances, such as carbon,
silicon, or sulphur, may also be present in the metal. If such
a metal be now poured, the outer skin of the casting will
solidify first, and then a considerable pressure will be set up in
the liquid material enclosed within it ; this pressure facilitates
the reaction (otherwise very sluggish) between the oxygen
compounds and whatever carbon, silicon, sulphur, and the like
may be present. The result is that gases are formed which
have no means of escape, and therefore tend to produce cavities
in the metal. Then, again, this metal has the power of dis>
solving other gases also, as, for example, hydrogen and nitrogen*
If a small quantity of aluminium be added to such a metal
either in the casting ladle before pouring, or aflerwards in the
ingot moulds, the substances which give rise to the difficulties
above described are in part reduced and in part separated. The
amount of aluminium added to cast iron and cast steel is always
smaller than would correspond to the amount of ferrous oxide
Digitized by VorOO^ It!
ALUMINIUM. 165
to be reduced. But as not only this oxide but other impurities
also are simultaneously removed, it is evident that the alu-
minium is able to diffuse very rapidly through the iron, and
in so doing to exert a powerful reducing action on the oxygen
compounds with which it comes in contact; and, further, it
^eems that the heat generated by this reaction expedites the
ireactions which otherwise would have gone on comparatively
slowly after the outer crust of the metal had solidified. In this
way the cast metal is purified so rapidly before the commence-
ment of solidification, that the time required for solidifying is
reduced, and, in consequence, any tendency to the separation or
liquation of constituent alloys is lessened, and the homogeneity,
density, and tensile strength are increased. In the iron foundry
the amount of aluminium used is commonly 500 to 1,000 grms.
[I'l to 2-2 lbs.] per ton ; in the open-hearth process it may be
50 to 150 grms. [If to 5^ ozs.] ; and in the Bessemer process for
formal runs 80 to 200 grms. [2^ to 7 ozs.], or, for under-blown
metal, more per ton. If too much aluminium be added the
tendency of the ingot to "pipe" will be increased, whilst a
moderate addition has the opposite effect. An excess of alu-
minium also causes a separation of carbon.
In the so-called " galvanising " (zinc-coating) process, aikd in
the brass foundry the rdle played by aluminium is rather that of
a reducing medium or a means for checking the oxidation of the
charge than that of an alloying agent. Hunt states that 500
grms. [1 lb.] of aluminium in the form of a 5 to 10 per cent,
aluminium-zinc alloy is used per ton of zinc, whilst, according to
Kichards, an addition of 2 kgrms. [4^ lbs.] of an alloy of 4 per
cent, of aluminium, with 96 per cent, of zinc, suffices for a bath
of 10 tons of the latter metal.
Finally, aluminium is used in the Goldschmidt process for the
separation of difficultlv-fusible metals in the molten condition
by the reduction of their oxides by its means {cf. chromium,
manganese, <bc.). Goldschmidt has also introduced a process by
which refractory metals, such as iron, can be obtained at very
high temperatures and applied especially to welding and solder-
ing. In the case of iron, a mixture of oxide of iron and
aluminium is lighted, and the resulting reduced iron, charged
with the whole of the heat generated by the reaction, is allowed
to run on to the joint to be soldered or welded. It is obvious
that this method of obtaining high temperatures is applicable
under conditions in which other means oi heating could only be
applied with great difficulty, such, for example, as the welding of
railway rails, or the repair of machine parts, without dismounting
the whole.
Frioe of Aluminium. — This chapter may be concluded with
an interesting table, taken from the publications of the Frank-
fort Metallurgical Society and of the Metall-GeseUachaft, showing
Digitized by V^OO^ It!
166
ELECTRIC SMELTING AND REFINING.
the variations in the price of aluminium since it was first
produced on a manufacturing scale.
Year.
Approximate Price.
Manufacturers.
Marks per
Shillings
Kilo.
per Lb.
1855
Deville, at Glaci^re, .
1,000
455
1856
,, ...
300
136
1857
Morin, at Xanterre, .
240
109
1857-1886
Merle & Co., at Salindre, .
100
45-6
1886
Hemelingen, ....
70
31-S
1888
Alliance Aluminium Co., .
( Aluminium-Industrie- Aktien-
47.50
21-6
1890, Feb.
Gesellschaf t, at Neuhauaen ; \
[ and other H^roult Works,
27.60
12-5
1890, Sept.
1891, FeV
ff t
15.20
6-9
9t t
12
5-5
1891, July
>) f
8
3-6
1891, Nov.
ft f
5
2-3
1892
>* 1
5
2-3
1893
)l f
5
2-3
1894
f 1 f
4
1-8
1895
tt »
3
1-4
1896
}) i
2.60
1-3
1897
** y
,
2.50
11
1898
If 1
2.20
1-0
1899
)) f
2.10
0-95
1900
If »
2.10
0-95
CHAPTER II.
THE CERITE METALS.
CERIUM, IiANTHANITM, FRASEODIDYMIITM,
IV^ODIDYMIUM.
Ooourrenoe in Nature. — These metals are found in nature as
silicates in cerite, and as phosphates in monazite. Since the
compounds of the rarer earth metals, especially those of thorium,
have been used in the manufacture of incandescent gas mantles
for lighting purposes, the compounds of the cerite metals (which
are contained in considerable quantities in these metals) have
become more readily available for use. The separation of the
compounds of metals of the cerite group by chemical means is
still very costly, so that, in the later experiments in the extrac-
Digitized by V^OO^ It!
THE CERITB MBTALS. 167
tion of the metals, the mixture of oxides or salts left from the
manufacture of incandescent materials has been submitted
direct to reduction or electrolysis without any attempt at
separation previously.
PropertieB of Cerium. — Cerium, Ce, the best-known metal
of this group, has the atomic weight 138, and the specific gravity
6*73. It has an iron-grey colour, is soft, being somewhat harder
than lead, malleable, and readily rolled. Its fusing point is
about 800* C.
Most noteworthy is its power of combining with the heavier
metals, such as copper or iron, to produce dense alloys. In solid
pieces it offers considerable resistance to atmospheric influences ;
but, on heating the fragments, they exhibit the various temper-
colours of polished steel. Finely-powdered cerium, on the con-
trary, oxidises very rapidly in the air, and, on filing the metal
or shaving it with a knife, the detached filings or scrapings take
fire and burn with a brilliant light. Fine wire made from the
metal burns with a brilliancy exceeding even that of magnesium.
Cerium in the form of powder causes only a slow decomposition
of water when introduced into it, but the presence of salts dis-
solved in the liquid, induces a very lively attack. This property
should be noted in connection with the production of cerium,
and with the passibUityy which may easily arise, of obtaining the
metal in pidvenUent form, owing to the use of an electrolyte ait too
low a temperatwre,* It dissolves very easily in diluted acids,
but only to a slight extent in cold concentrated sulphuric or
nitric acids. Cerium reduces the oxides of most metals and
metalloids, which is a property worthy of remark in regard to its
applications to metal refining and the making of alloys.
Lanthanum, praseodidymium, and neodidymium agree with
ceiium in very many of their properties.
1. EXTRACTION OF THE METAL BY
ELECTROLYSIS.
Eleotrolytic Deposition of the Cerite Metals. — ^The cerite
metals appear to form a group intermediate between those of
magnesium and aluminium, in regard to the behaviour of those
salts which would be likely to come into use for electrolysis. It
is well understood that none of the chlorides of these metals can
be obtained in the anhydrous condition by evaporating their
solutions in water, since decomposition always occurs on drying.
But, as in the case of magnesium chloride, if a chemically
equivalent quantity of the chloride of sodium or potassium,
* This is an exact reprint from the previous edition of this work.
Muthmann, Hofer, and Weiss have, however, re-discovered, in the year
1901, that for the separation of cerium in the molten condition the electro-
lyte must be hot, and in the neighbourhood of the cathode very hot.
Digitized by V^OOQ IC
168 ELECTRIC SMBLTINQ AND REFININO.
together with a little ammoniam chloride, be added, the sola-
tion of the cerium, lanthanum, or didjmium chloride may be
evaporated to dryness, and the dry residue may then be fused
without decomposition. The melt then contains comparatively
easily fused double chlorides of the cerite and alkali metals, and
will be found to conduct the electric current well. But although
it may have been easy to obtain the metals of the magnesium
group, magnesium and lithium, almost absolutely pure, either
by direct electrolysis or by electrolysis followed by fusion, it is
not safe, with the methods of production hitherto described, to
rely too much on the purity of the separated metal, if its re*
duction has been effected in quantities somewhat greater than
would be possible in the small porcelain crucibles of the labora-
tory. It is, indeed, very improbable that fiunsen, Hiilebrandt,
and Norton, who were the first to reduce the cerite metals
by electrolysis, obtained a product that was free from iron.
They employed the following method, which was devised by
iunsen *: —
The decomposing vessel in which the electrolysis of the molten
chloride was to be accomplished was arranged after the fashion
of a Grove's element The outer cell, which in the Grove's
battery contains the zinc plate and sulphuric acid, is here an
ordinary Hessian crucible of about 100 c.c. [34 fl. ozs.] capacity,
filled with a fused mixture of equivalent weights of sodium and
potassium chlorides, in which a cylinder of thin sheet-iron serves
as positive electrode in place of the zinc of Grove's cell. The
cylinder is 5 cm. [2 ins.] high, and 2*5 cm. [1 in.] across internally,
and terminates in a strip which serves as a conductor, and must
not be either soldered or riveted in place. Within the cylinder
is a clay cell of the best quality, 9 cm. [3^ ins.! high and 2 to 2*6
cm. [I to 1 in.] wide, in which is placed the chloride to be
decomposed. The negative electrode is immersed in this to
about two-thirds of the depth of the cell ; the electrode consists
of a thick iron wire, the end of which is filed down somewhat
thinner, and round its end is twisted a piece of iron wire about
as thick as a horse-hair, which projects some 15 mm. [^ in.]
beyond the stouter piece to which it is attached. A piece of a
clay pipe-stem is now drawn so far over the thicker wire, that
only the fine wire t at the end projects out of the clay and comes
in contact with the fused chloride that is to be reduced.
In the reduction of such chlorides as are easily converted into
oxides by the action of water vapour, the fusion must never be
effected by means of a gas flame. Even in the heating of
chlorides that are less readily decomposed, it is better to avoid
the use of gas flames, since the water vapour that they evolve
*Poffff, Ann., 1850, vol. dv., p. 633.
t A piece of this wire, 1 cm. long, weighs about 4 mg. [1 in. weighs
about 1^ grains.]
Digitized by VjOOQ IC
THB CSBITB METALS. 169
is very liable to cause re-oxidation of the already reduced metal.
The charcoal that is used to melt the electrolyte in these cells,
therefore, must be thoroughly glowing, and should have given
off all the hydrogen that it contained before starting the ex-
periment. For the same reason, the chloride that is to form the
electrolyte must be very thoroughly dried, and must then be
heated in a platinum crucible with sal-ammoniac until the bulk
of the latter salt has been expelled. It must be stored in closely
stoppered bottles, and be guarded most carefully against the
re^absorption of moisture. Finally, when the chloride is melted
for the experiment, the contents of the clay cell are covered with
a layer of powdered sal-ammoniac, which has previously been
heated, and this salt is replaced as fast as it volatilises.
The yield of metal and the size of the globules obtained
depend upon the temperature at which the fused chlorides are
submitted to the action of the currents. If the clay cell be
raised to a temperature exceeding the fusing points of the salt
under electrolysis and the metal that is to be separated, the
drops of metal which form upon the surface of the negative
electrode, fall to the bottom, and are there for the most part re-
oxidised at the expense of the silica in the clay walls of the cell.
The addition of fuel and the supply of sal-ammoniac are therefore
so regulated that the upper part of the salt in the clay cell
remains solid, while the lower part around the negative electrode
is in a semi-solid or pasty condition. The metallic particles thus
increase in size without sinking through the pasty mass, and
may even grow into globules the size of a hazel nut if the ex-
periment be carefully performed The electrolytic decomposition
should be started only when the melted salt is in the proper
condition, because otherwise the reduced metal is liable to
separate in a pulverulent form, and to mingle with the contents
of the clay cell, so that the formation of larger metallic globules
would be prevented.
The success of the reduction depends not only upon the tem-
perature of the bath, but also upon the absolute intensity of the
current employed. Four large carbon-zinc elements suffice for
the experiment. The clay cells of such elements should contain
250 C.C. [9 fl. ozs.] of nitric acid; the carbon rods should be 21
cm. long by 2*5 cm. wide by 4*5 cm. thick [8J ins. x 1 in. x 1| ins. J,
and the total available area of the zinc surface surrounding the
clay cell should be 590 sq. cm. [90 sq. ius.].
The author's experiments in the reduction of these metals by
electrolysis, conducted in the year 1888, only extended to the
treatment of a mixture of their chlorides. Hillebrandt and
Norton have apparently adhered too carefully to the process
devised by Matchiessen for the separation of strontium. It is
not necessary in this case to use a current-density of the grade
that is required for the separation of the alkaline-earth metals.
Digitized by V^OO^ It!
170 BLECTRIC SMELTING AND REFIKIK6.
The double chlorides of the cerite and alkali metals above
described may quite easily be electrolysed in an iron crucible,
which either serves as (or is in direct electrical connection with)
the cathode. (See MagnesivAn^ Figs. 9 to 12). In such an
experiment, with a cathode area of 500 sq. cm. [77*5 sq. ins.], a
current of 50 amperes and 6 to 7 volts should be used. Cerium
separates out first in the molten condition if the electrolyte be
suihciently hot, otherwise in a pulverulent form, as indicated in
the previous edition of this work.* Ic is by no means impossible
that the metals of the cerium group might be separated from
one another by fractional electrolysis, since lanthanum and the
didymiuin metals tend to deposit after cerium, even if they do
not come down until the whole of the cerium has separated out.
The didymium metals are less readily maintained in the fluid
condition than is cerium, and they usually deposit in the pulver-
ulent condition.
Stookem's Experiments in the Beduotion of Cerium. —
These experiments were undertaken in 1900 by L. Stockem, who
was at that time a pupil of Dr. fiorchers. He repeated the ex-
periments of Hillebrandt and Norton with a modified form of
apparatus ; he then proceeded to electrolyse the chloride — and
also the oxide, dissolved in chlorides and fluorides (both of
cerium itself, and of aluminium and the alkali metals) after the
manner of the Heroult process — using the current employed for
electrolysis to maintain the temperature of the bath; and finally,
he employed an externally heated apparatus with the object of
improving the process already described by Borchers.t An
experimental apparatus similar to that adopted by Borchers for
aluminium rciluction was used for the electrolysis of electrically
fused salts ; but instead of the ordinary cooled carbon blocks,
an iron wire terminating in a fine point was employed as
cathode. This apparatus yielded molten metal at first, and
the cathode- wire was but slightly attacked. Soon, however, the
area of the cathode surface had increased so much that the wire
was no longer heated to a temperature above the melting-point
of the metal ; and the latter then separated out only in the
form of a powdery deposit. Hence, Stockem reverted to the
use of externally heated melting vessels, and thus ensured that
the charge, or at least the cathode and that portion of the
charge immediately surrounding it, was hot enough for the
metal to remain fused throughout the operation. A conical
vessel tapering to a point was used in place of the cylindrical
and slightly tapering melting vessel commonly employed for
magnesium reduction, because a higher current-density could
•Borchers' Electrometallurgies 2nd Ed., 1895, p. 165; Electric Smeltitig
and Refining, 1897, p. 166.
t Diploma work of L. Stockem, presented at the Technical High School
at Aachen. June 28, 1901.
Digitized by V^jOO^ It!
THE CERITE METALS. 171
then be applied at the cathode, whilst the anode could be
enlarged and the current-density reduced at that point. The
upper part of the melting vessel was provided with a cooling*
jacket lined with a clay or magnesite cylinder, so that there
shoiQd be no dauger ot iron dissolving from the walls of the
apparatus into the fused salts and, by subsequent reduction,
passing into the cerium.
Muthmann, Hofer, and Weiss' Experiments. — A paper^
by Muthmanu, Hofer, and Weiss on the reduction of the metals
ot the cerium group gives an account of a very noteworthy
advance in the utilisation of the oxides of these metals, even if
the views expressed in the introductory part cannot be accepted
in their entirety. The authors have described the experiments
with fused electrolytes, such as are commonly made by students
in the first week of their studies in the laboratories of the
Aachen Technical High School. Agaiu, they have disregarded
earlier work in the same field — and have thus rediscovered facts
long since established, although the recognition of such work
would in no way have detracted from the service rendered by
the authors in this research.
They found Bunsen's apparatus too small and complicated for
their purpose, whilst the Borchers apparatus tried by them was
discarded as being insufficiently compact, as requiring too much
material, and as being too inconvenient in use. They do not,
however, say which form of Borchers' apparatus they employed;
but as in a later part of their account they refer to certain earlier
experiments in which they electrolysed cryolite, Borchers assumes
that they must refer to the form of experimental furnace which
he had described in 1898 as being constructed for this purpose, f
He therefore places side by side drawings (Figs. 102 and 103)
of his furnace and of that used by Muthmann, Hofer, and Weiss,
and found by them to be satisfactory. A single description
will suffice for the principal parts of each form of construction,
namely : — A double-walled water-jacket, made of sheet copper,
surrounds the melting space, which is almost closed at the
bottom by the carbon cathode. The cathode is surrounded
by non-conducting material, which fills up the space between
it and the copper walls of the vessel. The anode is a vertical
carbon rod projecting downwards into the furnace from above.
Such a furnace is used by new students in the author's labora-
tory for their first experiments in the electrolysis of fused
materials. The authors' apparatus appears, then, to be only
a smaller form of Borchers* furnace. So far as the electrolysis
of alumina dissolved in cryolite in this type of furnace is con-
cemedf the larger the size of the apparatus, the more is the
use of small carbon surfaces within a cooling-jacket to be recom-
* Liebig^s AmvoUen der CJiemie, 1902, vol. cccxx., p. 231.
t Zeitechrijl/Hr Elektrochemie, 1898, vol. iv., p. 523,
Digitized by
Google
172
ELECTRIC SMELTING AND REFINING.
mended on the groands both of economy and safety. Borchers
ha8 never ^ however^ recommended the use of cooled cathodes for the
eleetrolf/ais of cerium cJUoride, as the authors appear to believe.
On the contrary, in the last edition of this book* he spoke "of
the possibility, which may easily arise, of obtaining the metal in
pulverulent form, owing to the use of an electrolyte at too low a
temperature." Yet the authors draw the same, conclusion from
their own experiments,
whilst disregarding the
statement made by
Borchers in 1895. The
following results ob-
tained by Muthmann,
Hofer, and Weiss are in
all respects noteworthy.
Treatment and Purifi-
cation ^ eitpecicUly of Pho8-
phatic Ores and Crude
Oxide. — Oerium dioxide
was the material avail-
able for the author's ex-
Fig. 102. — Section of Borchers' furnace.
Fig. 103.— Section of Muth-
nian, Hofer, and Weiss*
furnace.
periments, and the diiiiculty in converting this compound into
cerous chloride is well known to all who have worked with
it. The substance is not attacked by nitric or hydrochloric
acid. In order to dissolve small quantities in hydrochloric acid
for analytical purposes potassium iodide is added, which on
evaporation facilitates the solution of the oxide with evapo.
* See p. 167 above and footnote.
Digitized by LjOOQIC
THE CEBITS MBTALB. 173
ration of iodine. Obviously, however, this method is not
applicable on a large scale, not only because it is too costly, but
because it leads to the introduction of too much alkali-metal
chloride into the solution. The only way out of the difficulty was
to dissolve the dioxide in strong sulphuric acid, reduce it with
aloohol, decompose the resulting cerous sulphate by digesting
it for several days with soda solution, and dissolving the car-
bonate in hydrochloric acid after washing it with the greatest
possible care. Every trace of sulphuric acid had to be com-
pletely removed, since the smallest proportion of sulphate
interferes most seriously with the electrolytic process. It was
for this reason that Bunsen in his experiment precipitated the
sulphuric acid with barium chloride.
Of the crude oxides used by the authors, only those obtained
from cerite were soluble in hydrochloric acid. The inexpensive
material from the by-products of thorium manufacture was
scarcely soluble in hydrochloric acid, and only incompletely so
in strong nitric acid, even after repeated evaporations. This is
chiefly due to the material containing a considerable proportion
of phosphoric acid, which, considering its low cost, is scarcely
surprising, since all of these materials are obtained from strongly
phosphatic sources. The complete removal of the phosphoric
acid, which must be ensured before electrolysis, is even more
troublesome than that of sulphuric acid. Moreover, the con-
version of nitrate solution into pure chlorides, when prepared
with the strong acid, is a tedious process.
All these difficulties are removed by the authors, by first
converting the mixture of insoluble oxides and phosphates into
carbides, a process which gives excellent results, and is especially
convenient in cases in which there is a cheap source of energy
available. When the phosphates are electrically heated with
carbon they are converted mainly into carbides ; a portion, how-
ever, becomes phosphide, but by the addition of acid this may
readily be converted into the corresponding cerium salt with
evolution of hydrogen phosphide. The resulting solutions have
on no occasion been found to contain phosphoric acid, so that
this acid is safely and completely eliminated by applying a
process of electric smelting.
For the production of the carbide, of which a further account
will be given later, the authors used a graphite crucible to serve
both as melting vessel aud cathode, with a central carbon anode-
suspended from above, exactly as in the arrangement employed
by H^roult in his first experiment with aluminium.
After describing the process of carbide production and expres-
sing their views as to the existence of several carbides of cerium,
the authors point out that when the cerium carbide had been
dissolved in the smallest possible quantity of hydrochloric acid
and the resulting solution had been filtered, the iron, which
Digitized by V^OOQ IC
174 ELECTRIC SMELTING AND REFINING.
almost always passes into the compound in the process of smelt-
ing, had to be removed. To this end the solution was carefully
neutralised, and the resulting liquid, which contained about
120 grammes of cerous chloride to the litre [approximately a
12 per cent, solution], was carefully treated with a very dilute
and finely-divided spray of ammonia blown into it with the aid
of steam, until a test portion gave no further reaction after
filtration. The iron was first brought to the ferric state by
means of bromine water, so that it could be completely separated,
as above described, with very slight loss of material. A quantity
of ammonium chloride, equal to that of the carbide used, was
now added to the clear filtered liquid, which was then evaporated
over a gas furnace, and finally dried with constant stirring.
The resulting material was then heated to a dull red heat in a
large platinum dish, again with constant stirring, in small por-
tions of 50 to 100 grammes [If to S^ ozs.]. The ammonium
chloride escaped in dense clouas, and the cerous chloride was
finally obtained in the form of a fine pure white powder, which,
on stirring, behaved almost like a fluid. It was never allowed
to cool in the air, but was transferred into globular flasks while
still hot, and was there protected from the moisture of the air
by means of stoppers provided with tubes containing phosphorus
pentoxide.
The anhydrous neodidymium chloride used for the production
of that metal is prepared similarly, except that the preliminary
treatment in the electric furnace is unnecessary by reason of
the ready solubility of neodidymium oxide in hydrochloric acid.
Muthmann and Stutzel * previously, and Camille Matignon f
more recently, have described the properties of the anhydrous
neodidymium chloride. A short description may be given of the
purification of the neodidymium materials, as carried out in the
laboratories at Miinich by R. Bohm and L. Weiss. After a
portion of the lanthanum and praseodidymium had been removed
in the usual way by crystallising out the double ammonium
nitrate, the remainder of the lanthanum was removed as far as
possible by fiunsen's sulphate process. The material then still
contained some praseodidymium and large quantities of earths
of the yttrium group. The greater part of both these impurities
were removed by the chromate method described by Muthmann
and Bohm. I The first fractions to be precipitated were rich
in praseodidymium, the next in neodidymium, whilst the last
fractions contained most of the yttrium earths. Hence the
neodidymium was concentrated chiefly in the middle fractions ;
55 chromic acid fractionations, corresponding to about 800 pre-
cipitations, were carried out. This treatment with chromic
* BerichU der deuUcJien chem, OeseUschaft, 1899, vol. xxxii., p. 3413.
t Coniptea Bendua, 1901, vol. cxxxiii., p. 289.
t BerichU d. d, chem, Oeaellsch,, 1900, vol. xxxiii., pp. 42-49*
Digitized by V^jOOQ IC
THE CERITE METALS. 175
acid was continued until about 300 grms. [10^ ozs.] of material
almost free from praseodidymium were obtained, but there were
always appreciable quantities of gadolinium, yttrium, erbium,
and terbium left after this. These yttrium earths were next
separated by the usual potassium sulphate process, and the
material was afterwards subjected to two other fractionation
methods to ensure complete purification. The oxalates were
precipitated from hydrochloric acid solution, leaving the colour-
less yttrium earths in the liquid ; and, lastly, the material was
divided up into 79 portions by fractionating eight times with
very dilute ammonia solution, whereby colourless yttrium
earths, terbium, erbium, samarium, neodidymium, and praseo-
didymium were successively precipitated. Those fractions which
exhibited identical spectra were mixed, and a small portion
divided into 13 parts by nitrate treatment. The first and last
fractions were found to be exactly alike, which afforded proof
of the purity of the neodidymium. The material so obtained
yielded an oxide of pure blue-grey colour without any shade
of yellow or red, such as is exhibited by those samples which
still contain traces of heavy yttrium earths or of praseodi-
dymium. The spectrum agreed exactly with that described by
Schottlander ; * the line X 468'9 was especially sharp and clear,
which was a proof of the great purity of the material.
Extraction of the Metal. — After the three authors above re-
ferred to had determined by their experiment in the electrolysis
of the pure chlorides (especially of cerium and neodidymium)
that a high temperature was necessary for the separation of the
cerite metals in the molten condition, they proceeded to reduce
an alloy of the metals of the cerium group by treating their
mixed chlorides, which are readily obtained. In doing so they
worked on a somewhat larger scale than heretofore, and con-
structed the very practical form of furnace shown in longitudinal
section, plan, and elevation respectively, in Figs. 104, 105,
and 106.
The melting space or crucible was oval in shape, the shorter
axis measuring 7 cm. [2| ins.], the longer axis 11 cm. [4^ ins.],
and the height 7 cm. The furnace was capable of being exter-
nally cooled throughout the whole of its height, and was
attached by a clamp to a rod, which also carried the holders
for the cathode and the anode. These latter could be moved
upwards or downwards at will. A special device was added to
ensure that the charge in the neighbourhood of the cathode was
heated to a sufficiently high temperature. This end was attained
with the aid of two small resistance-pieces, each connected to
a pair of carbons, connected in parallel in an alternating-current
circuit. The alternating current was chosen for heating because
there would, in this case, be no electrolysis at the ends of the
* Berichte d, d. chem, OesdUeh., vol. xxv., p. 669.
Digitized by LjOOQ IC
176
ELECTRIC SMELTING AND REFINING.
carbons such as would doubtless occur if continuous current
were used, an action which would interfere with the process.
The incandescent rods were connected to the thicker carbons in
such a way as to ensure a good joint. The thick carbons were
held in iron clamps, which could be fixed at any required height
with the aid of supports sliding up or down in hollow pillars.
The current connections were made by means of the clamps
marked ± and T in the figures. Both the cathode and the
alternating-current carbons were insulated from their supports
by means of asbestos cord or asbestos paper, and the whole
Figs. 104, 105, and 106.— Furnace for extraction of cerium metals.
apparatus was mounted on a heavy oak stand. An arrangement
for tapping the charge has been added to the furnace, but
hitherto the authors have never required to use it.
In order to test- these alternating-current furnaces, as the
authors have termed them, a trial run was made on aluminium
extraction. The crucible was filled with cryolite, and the
alternating and continuous currents were switched on simul-
taneously. Each of the three carbon pencils was 5 mm. [^ in.]
thick, and 25 mm. [I in.] long; the anode carbon had a diameter
of 40 mm. [1 J ins.], and the cathode 30 mm. [1^ ins.], whilst the
Digitized by V^OO^ It!
THE CBRITE METALS. 177
carbons for the alternating current were 15 mm. [4 in.] thick.
The capacity of the cmcibley with the electrodes dismounted,
was 400 C.C. [24^ cub. ins.], and the vessel was capable of con-
taining 700 grms. [1^ lbs.] of molten cryolite. The current-
conditions during the melting period were as follows : —
Ck>ntiniioiiB current, 75 ampere* at 10 volts.
Alternating ,> 80 „ 8 „
The fusion took place satisfactorily, but it was necessary, at
the end, to sprinkle potassium-sodium chloride on the surface
and so to reduce the melting point very considerably, in order
to remove a solid crust formed there. After about twenty
minutes the vessel was seen to be filled with a clear transparent
liquid, in which the strongly incandescent alternating-current
pencils were clearly visible. The continuous-current pencil was
then removed and electrolysis commenced with a current of
20 amperes and 50 volts. After the addition of fresh material
the current-strength was increased to 25 amperes, iu order that
it might range between 20 and 32 amperes according to the
position of the anode throughout the progress of the electrolysis.
The potential difference was gradually raised from 60 to 65 volts.
Electrolysis took place extremely quickly, and without any
accident. After four hours the current-circuit was broken and
the charge, after cooling, removed from the vessel. In order
to accomplish this conveniently, the alternating-current carbons
were removed, being bored out of the mass with the aid of the
lathe. On breaking up the pure white mass within the crucible,
two large globules of aluminium were found separated out on
the surfaces of the cathode facing the alternating-current pencils.
Their united weight was 23 grms. [355 grains] ; but besides
these there were ten smaller globules of from 2 to 4 mm. [0*08
to 0*16 in.] in diameter, weighing 2 grms. [31 grains] in alL
According to the laboratory journal it appeared that 91 ampere-
hours had been used, so that the gross yield was 82*2 per cent,
of that theoretically possible.
It was now chiefly necessary to ascertain whether the furnace
was suitable for the reduction of the cerium metals, and for this
purpose the mixed chlorides produced from monazite of known
composition, were submitted to electrolysis. The charge was
melted exactly as in the case of cryolite, and after twenty-
seven minutes the continuous-current pencil was removed. The
measuring instruments at the commencement of electrolysis
showed a current-strength of 105 amperes and a pressure of
15 volts. After a quarter of an hour the volume of current
increased to 125 amperes, and the pressure fell to about 10 volts.
The very brisk evolution of chlorine that resulted led to the
formation of bubbles; and there was also a crust of solid
material formed ; but this could be avoided by sprinkling the
forface with the double chloride of potassium and sodium.
^2 I
Digitized by V^OOQ IC
178 ELECTRIC SMELTING AND REFINING.
Unfortunately, after about an hour and a half, two carbons of
the resistance used were burnt out and had to be cut out of
the circuit, and the current- volume could only be kept constant
with the aid of the shunt-regulator. In consequence of this
a strong sparking began which would, of course, have led
to the burning of the brushes and a rapid destruction of the
commutator. The only course left, it the experiment was to be
continued unchecked, was to increase the internal resistance of
the charge. This was effected by adding 55 grms. [2 oz&] of
completely dehydrated barium chloride to the 1*15 kilogrms.
[2^ lbs.] of fused mixed chlorides in the crucible. The desired
result was at once achieved ; there resulted a steep potential-
gradient between the anode and the fused charge, characterised
by the formation of small arcs and the " singing " of the anode.
The voltmeter reading rose to 33 volts. After a short time the
pressure was still further increased to 42 volts by the addition
of 10 grms. [154 grains] more barium chloride, and this pressure
was maintained for the remaining few hours of the run. At
first the current- volume decreased to 60 amperes, but was raised
to 85 amperes by gradually lowering the anode. The volume of
current gradually fell, in the end to 55 amperes, in proportion
as the anode (to avoid all danger of short circuit) was raised,
and as the anode area was, in consequence, reduced; but the
pressure remained constant.
The alternating-current pencils, were not altered during the
whole of the 6^hours period of electrolysis, during which the
current was constantly maintained at 98 amperes and 8 volts.
The use of the barium chloride in the electrolysis of the
chlorides of the cerite metals is worthy of special attention.
The possibility of varying the pressure and strength of the
current by the addition of a small quantity of this salt is
remarkable. If there should be a marked foaming or spitting
of the charge, such as often occurs without any recognisable
cause, the sprinkling into the bath near the anode of just so
much barium chloride as will cover the point of a knife, at once
causes the decomposition to proceed more quietly in consequence
of the increase in the difference of potential so induced. If the
barium chloride be introduced very cautiously, a point may
easily be reached at which, in a short time, the pressure varies
markedly, as may be seen by the rapid oscillation of the needle
of the ampere-meter. This condition may, at will, be maintained
for hours by introducing, alternately, first mixed chlorides and
then a few grains of barium chloride. The phenomenon appears
to be due to the charge not wetting the anode, if a small pro-
portion of barium chloride be present, whilst the fused chlorides
of the cerite metals, if mixed only with alkali-metal chlorides,
are capable of wetting it just as water would. The addition of
barium chloride has no effect upon the composition of the separ-
ated metal, which never contains a trace of metallic barium.
Digitized by VjOOV It!
THE CERITE METALS.
179
After 6^ hours the conductant metallic product of electrolysis
had risen in the crucible to such a height that a short circuit
took place close beneath the surface of the charge. The current-
circuit was then broken, and the apparatus was cooled and taken
to pieces. A single regulus weighing 170 grms. [6 ozs.] was
found, a fact which speaks well for the serviceableness of the
apparatus. This end could only have been attained, however,
Fig. 107. — Borchers' furnace for cerite metals.
by^the application of the heating effect of the alternating current,
for, in earlier experiments made without this, not even a trace
of fused metal was ever obtained. It is further evident that it
is of great importance to have a high temperature in the neigh-
bourhood of the cathode, and that the arrangements for cooling
the cathode found in many types of apparatus are useless for
experiments such as those just described.
Digitized by VjOOQIC
180
ELECTRIC SMELTING AND BEFININO.
The current-efficiency was not good. The deposition of 170
grms. [6 ozB.] of metal after an expenditure of 280 ampere-houra
corresponds to a current-efficiency of 35*7 per cent., whilst, in
the same apparatus, when depositing aluminium it was 82*2 per
cent. The explanation of this is to be found in the convection
which always takes place with the cerite metals, in consequence
of metal in a finely-divided condition passing into the fused
mass. The various metals appear to behave very difierently in
this respect. Experiments made by Lorenz, and others made
by Mnthmann and his collaborators, show that the phenomenon
is very marked in the case of lead, whilst it is scarcely observ-
able with aluminium. The cerite metals occupy an intermediate
position between the two. Apparently, the softer the metal
the more subject is it to this action. It is probable that certain
other types of furnace used by Borchers, in which the cathode
is heated electrically (such as that shown in Fig 107) might
be adapted for this work.
2. REDUCTION PBOCESS.
As stated in the previous edition of this work, the oxides of
the cerite metals can readily be reduced by electrically-heated
carbon, notwithstanding statements to the contrary in chemical
text-books. But, as in the case of the oxides of the alkaline-
earth metals, a metallic product is not obtainable in this way, by
reason of the strong tendency of cerium to form carbides and
nitrides. The carbide appears to mix in all proportions with
the oxides, as is proved by the following experiment conducted
by L. Stockem * in Borchers' laboratory.
For the experiment a weighed quantity of cerium oxide was
kneaded to a thick paste with tar diluted with benzine, and waa
then boiled in a crucible furnace. This process was repeated
again and again until the most intimate mixture possible of the
oxide and carbon resulted, in proportions corresponding to the
formula Ce^Og +30. The mixture was then heated to the
reduction-temperature of cerium oxide in the electrical resis-
tance furnace just described. The thoroughly-fused mixture of
oxide and carbide so obtained had the following composition,,
as found by analysis : —
Per cent
Per cent.
CeC, . , . =
Ce,d, . . =
36-69
63-25
3710
62-91
99-94
10001
' See Note, p. 170.
Digitized by LjOOQ IC
THE CEBITS METALB.
181
It was yellow in colour, and its fracture was crystalline;
when it was brought into contact with water or acid, it
liberated marsh gas freely ; and, when scratched with a knife, it
emitted sparks, but not nearly so briskly as is the case with
metallic cerium. This carbide gave a decided odour of ammonia
when exposed to moist air, a phenomenon that is much more
evident if a larger proportion of carbon is used for the reduction
of the cerium oxide, and if the heating is conducted in an atmo-
sphere of nitrogen. Although the odour of ammonia might lead
to the supposition that a large proportion of nitrogen was
present, there was really but little. The result, however, was
sufficiently noteworthy to encourage the undertaking of an
investigation into the conditions favourable to the formation of
nitrides; the results of this investigation will be given elsewhere
when completed. The experiment just described clearly shows
that it is hopeless to expect metal to be reduced by a reaction of
the carbide with the oxide of cerium.
FigB. 108 and 109. — Stockem's furnace for production of cerium alloys.
Bedaotion of Cerium Alloys. — It is well known that
aluminium oxide, when reduced by carbon, does not yield a
pure metallic product ; but, in spite of that, the brothers Cowles
succeeded in so reducing aluminium alloys by adding other
metals, or their oxides, to the charge. Similarly, L. Stockem,*
smelting a mixture of cerium oxide, copper oxide, and carbon in
the proportion that would be expected to yield an alloy contain-
ing 10 per cent, of cerium, obtained a metal containing 94*43 per
cent, of copper, and 5*5 per cent, of cerium. In a second
experiment, the proportions of the ingredients were so chosen
• See Note, p. 170.
Digitized by VjOOQIC
182 SLECTRIC SMELTING AND REFINING.
that an alloy containing 20 per cent, of cerium might theoreti-
cally be anticipated ; but, again, only half of the expected
proportion of cerium was obtained, the composition of the alloy
being Ca — 89*67 and Ce = 10-3 per cent. In each case the
greater portion of the unalloyed cerium was found to have been
converted into carbide. It was not possible to obtain an alloy
of iron and cerium in this way.
A furnace of somewhat different construction to those
previously used was found to give good results in these experi-
ments. An empty magnesite crucible (Figs. 108 and 109) was
placed as a resistance between the carbon rods carrying the
current. But as this material is a non-conductor when cold, it
was packed round with powdered hard-wood charcoal of large
grain and low conductance, and heated to a high temperature.
As soon as the crucible was thus raised to a bright red heat it
began to act as a conductor, and the temperature then rapidly
rose to that of a brilliant white heat by the action of the current
flowing through it. The mixture to be reduced was now placed
in the crucible, and the reaction occurred briskly and in a short
space of time.
Extraction of Cerium by Metallic Reducing Agents. —
The experimental substitution of aluminium and magnesium for
carbon, after the manner of the Goldschmidt process, was not
found to be a success. The excess heat of the reaction (which is
doubtless a possible one) is evidently insufficient to bring about
the desired result. But if a source of heat, such as the furnace
just described, be used, it is very difficult so to adjust the
temperature to that required for an adequate yield of the metal,
since the boiling point is obviously lower than the optimum
temperature for the reaction. The result of using calcium
carbide as a reducing agent with a mixture of the chloride and
oxide of cerium is equally unsatisfactory.
But although the pure metal cannot be thus got, it is possible
to obtain alloys of cerium by adding oxides or chlorides of other
metals to the charge to be reduced by aluminium (or, it may be,
by calcium carbide). Mixtures of the oxides of cerium and
copper with aluminium did not give exactly the expected result,
because a part of the added aluminium was found to have
alloyed with the cerium and copper reduced. The experiment,
however, at least proved the possibility of reducing cerium oxide
by means of aluminium.
Uses of the Cerite Metals. — Since the methods of prepara-
tion of the cerite metals are still quite in the experimental
stage, it is obviously impossible at present to say much as to
the industrial applications of these metals. The openings for
utilising them are, however, fairly numerous, especially as con-
siderable quantities of the cerite oxides may reasonably be
expected to come into the market as cheap by-products from
Digitized by VorOO^ It!
THE CEBITE METALS. 183
oerite and monazite in the manufacture of the mantles for
incandescent gas burners.
The metals and carbides would, doubtless, at a low price, be
used in the manufisicture of fireworks on account of the brilliance
of the sparks which they produced in burning. The light would
probably be useful also in photography. The readiness with
which the cerite metals form alloys, especially with copper, is
noteworthy, as are also the density, toughness, and relatively
great hardness of these alloys. The reducing power of cerium
approaches that of aluminium. It should be observed also that
a study of the formation of the nitrides of the cerite metals may
lead to the synthesis of other nitrogen compounds.
Digitized by VjOOQ IC
184
PART III.— THE HEAVY METALS.
CHAPTER I.
COPPER.
Occurrence of Copper in Nature. — The following are the
ores principally used for the extraction of copper: — Native
[metallicl copper; the oxides, red copper ore or cuprite, CujO,
and blacK oxide or tenorite, CuO, which is much less common ;
copper glance, Cu^S ; copper pyrites or chalcopyrite, CujS . Fe^^ ;
variegated copper ore or embescite, dCu^S . FegS, ; the salts,
blue vitriol, CuSO^ . bUfi ; malachite, Cu2(OH)2C03 ; and blue
carbonate of copper or azurite, Cu3(OH)2(C03)2.
But besides the above natural compounds, there are many
artificial products containing copper which may be used for the
extraction of that metal. Among these are mattes, produced
incidentally in the smelting of lead or nickel, speise, slags, and
alloys j of esi)ecial importance among the latter may be named,
dross from liquations and certain products from the refining
of lead.
The ores and the cupriferous by-products are usually first
treated so as to yield an impure metal, crude copper or black
coppery which is then treated for the preparation of marketable
metal. The following short sketch of the general metallurgy
of copper will assist the consideration of attempts, in many
cases unsuccessful, to replace some of the purely chemical or
metallurgical processes by electrolysis :— -
Digitized by VjOOQ IC
COPPER.
185
I. EXTRACTION OP CRUDE COPPER.
Smelting. Solution and Precipitation.
Available for all copper ores. Available for carbonates, sul-
phates, and oxides.
A. Solvents.
WcUer for the sulphate.
A. Enriching Pr
1. Oxidising roast. — Object: sep-
aration of part of sulphur.
2. Smdting for regiUus or mcUte . —
Object : combination of the copper
and a portion of the other metals
with sulphur ; removal of portion of
constituents other than copper in
the slag. Roasting (1) and smelting
for regulus (2) are united in : —
Kernel roasting^ in which the cop-
per sulphide liquates into the middle
of the fragments during roasting,
and
Pyritic smelting, in which the
sulphide is submitted to oxidising
smelting either without, or with a
little, lulditional fuel in the blast-
furnace.
3. Oxiditiing roast of regulua. —
Object as in (1).
4. Smelting Jor concentrated regn-
Ins. — Object as in (2).
The processes (3) and (4) are also
united sometimes in reverberatory
or converter processes.
B. Beoovery of Crude Metal.
Consisting of the reaction process,
so conducted in the reverberatory
furnace or converter that the oxides
react with undecomposed sulphides
to form SO3 and metals ; or the
reduction process, an older and less
used process, in which the regulus
was uead-roasted in order to re-
move the sulphur as completely as
possible, and the oxide was then
reduced to crude copper (black cop-
per), the substances other than
copper being removed as fully as
practicable by the euidition of silica.
The electrolytic treatment of regulus
has so far been unsuccessful.
Hydrochloric add for ores and
n etallurgical products containing
oxides, basic carbonates, and sul-
phates.
Sulphuric add and atmosphefi'ic
oxygen, acting simultaneously, for
the treatment of alloys rich in cop-
per, especially those containing the
noble metals. Copper thus passes
into solution as sulphate, the noble
metals with lead and other metallic
compounds remaining undissolved
in the form of mud.
Ferric and cupric salts, preferably
ferric chloride, which is lormed in
the treatment of ores containing
iron with hydrochloric acid in the
presence of air ; available for cuprite
(CujO), and, by prolonged action,
for sulphides that are readily
attackable.
B. Precipitation.
By metals of higher solution pres"
mire than copper, — Iron scrap is used
almost exclusively owing to its low
price.
By electrolysis. — ^Attempted with
but little success up to the present,
bv Siemens and Halske, and C.
Tloepfner, who endeavoured during
the precipitation of the copper to
regenerate the solution (ferric or
cupric salt) at the anode.
II. PURE COPPER.
Fumaoe Frooesses of Refining may be as follows : —
1. Oxidising melt, by which the substances other than copper
oxidise and are removed, partly by volatilisation (SO, . AsgO^)
and partly in the slag (metals). A part of the copper also is
oxidised to CngO, which dissolves in the pure copper.
Digitized by
Google
186 ELECTRIC SMELTING AND REFINING.
2. Reducing process for the reduction in the furnace of the
OujO produced and dissolved in stage 1 : — Product, malleable
copper.
Black copper containing the precious metals is purified by
electrolyeis, either direct or after partial refining by process ii., 1.
Anodes = black copper ; electrolyte = acidified aqueous solution
of copper salts ; cathodes = electrolytic copper.
Properties of Copper. — Copper (Cu* and Cu'; atomic weight
= 63*4, specific gravity = 8-94), when pure, exhibits a yellowish-
red colour on surfaces recently fractured. The structure of cast
copper is granular, but by hammering or rolling it becomes
fibrous. It is remarkable for great ductility, united with a
moderate degree of hardness and tenacity. Copper which has
become hardened by mechanical work may be made malleable
again by heating. Its conductivity for heat and electricity is
very high, and the latter property has therefore led to its general
use for conductors in electro-technical work. Shortly before
attaining its melting point (which is about 1,050° C.) the metal
becomes so brittle that it may be pulverised. In the molten
condition copper has a greenish colour, and it also imparts a
green tint to oxidising flames. The boiling point of the metal is
not far removed from the temperature of the oxyhydrogen flame,
and may perhaps be taken as approximating 2,000*" C.
When melted, copper has a great tendency to dissolve certain
gases (hydrogen, carbonic oxide, and sulphurous acid), which it
evolves again on solidifying. Certain metals (aluminium, nickel,
cobalt, zinc, cadmium, tin, lead, bismuth, the noble metals,
manganese, chromium, tungsten, molybdenum, and iron) are
also readily dissolved by it, as also are various metallic compounds
(cuprous oxide, sulphide, and phosphide, and the arsenides,
arseniates, antimonides and antimoniates of lead, bismuth, &c.).
The solubility of the gases and the compounds last named should
be remembered in connection with the melting of the copper in
the refinery, and in the casting of copper and its alloys generally.
The mutual solubility of copper and other metals is utilised in
the preparation of alloys.
Of the chemical properties of this metal the following may
be specially noted as bearing upon its extraction and useful
applications : — The polished surface of rolled or hammered (and
therefore dense) copper may remain unaltered in the air for a
long time ; but, in contact with acid substances in the presence
of air, the metal is very easily oxidised with the formation of
basic salts (e.^., verdigris). Or, again, at a low red heat, far
below its melting point, it becomes covered with a film or "sccUe'*
which consists of a mixture of the cuprous and cupric oxides.
In combining with sulphur it exhibits a far greater chemical
energy than does any of the other heavy metals, and this property
has a special interest in connection with the extraction of copper
Digitized by V^jOO^ It!
COPPER. 187
from its ores. Copper also unites directly with the other metal<
loids, excepting hydrogen, nitrogen, and carbon.
The principal solvents for copper are nitric acid, concentrated
sulphuric acid, and aqua regia. Hydrochloric acid and dilute
sulphuric acid only dissolve the metal when air or some other
oxidising substance is present ; on account of the low solution
pressure of this metal, hydrogen can never be produced by the
action of acids upon it, and it is for tliis reason that the addition
of an oxidising agent is required.
Copper enters into combination either as a monovalent (in
cuprous compounds) or as a divalent (in cupric compounds)
element In the salts of the oxygen acids it is almost always
in the divalent condition.
Electrolytic Beflnlng of Copper. — The foregoing sketch
has indicated that of all the attempts to apply electro-chemical
processes to the extraction of copper, only that of electrolytic
refining has found any wide practical application in metallurgy.
It will be best to commence with a study of this process on
account of the simplicity of the procedure connected with it.
EIiECTBOLYTIC REFINING OF COFFER.
Early attempts in the Electro-Deposition of Copper. —
Although the fact that copper could be separated from its
solutions by means of the electric current, was known through
Cruikshank's* researches in 1800, no successful application of
this observation could be made in the field of metallurgy until
after the invention of the dynamo. Following upon this, how-
ever, no time was lost in utilising the current for the purpose in
question.
Attempts, however, were constantly made to apply the cur-
rent to metallurgical purposes, and especially to the treatment
of copper ores, even at a time when there was no economical
generator of electricity available. Becquerel's work in the years
1835 to 1840, on the electro-chemical treatment of ores containing
silver, copper, and lead, will be referred to briefly under Lead,
As soon as Jacobi's discovery of electrotyping became known in
the year 1838, Smee recognised how important a part electrolysis
was destined to play in metallurgy, as may be gathered from
his work on electro-metallurgy. t Even in 1867, the year in
which Siemens first described his dynamo-electric machine,
Fateral published an account of a method for the precipitation
ol copper from cement waters, in which he wrote : —
* Nicholson's JoumaJ of Natural Philosophy ^ 1800.
t Smee's MectrO'inetallurgy, 1841.
X VerharuUungen dtr h. h. geo/oyischen Reichsaiu^taJt, 1867, No. 6 ; and
ZHngler^a polyieckiiisches Jotimalf 1867, vol. clxxxiv., p. 134.
Digitized by VjOOQ IC
188 ELECTRIC 8MELTINO AND REFINING.
Patera's Frooess. — "In a trough lined with guttapercha
plates, a rectangular clay cell was so cemented, that the two
walls that were in contact with the longer sides of the trough,
made a water-tight joint with them ; an empty space was thus
formed, bounded by the shorter sides of the trough, the free
walls of the clay cell, the bottom of the trough and that of the
cell. This space was filled with granulated copper, and through
it was passed the solution from which copper was to be recovered.
Within the clay cell were parallel iron plates, so soldered on to
a stout copper wire that there was a space of about four lines
between each pair of plates. This system of iron plates was
brought into electrical connection with the granulated copper by
means of a copper wire. The copper sulphate liquor was con-
tinually introduced on one side of the cell, and was run off from
the other through a glass tube. A solution of blue vitriol, with
a proportion of copper equal to that of the Schmollnitz mine
waters — viz., ^jj oz. per cb. ft. — loses more than half its copper
in traversing even the 1^ ft of space through which it has to
pass in this small apparatus. By passing the liquid through
two such cells in succession, or by leaving the solution in one cell
for only a short time, the extraction of copper will be complete.
** It appears, therefore, that this apparatus will answer all
requirements. The copper obtained in this way will be very
pure, the consumption of iron will correspond very nearly to the
equivalent of the copper separated, and the apparatus will be
very compact, and therefore easy to supervise. The only
difficulty tiiat I [t.^., Patera] observe is that a large quantity of
granulated copper must be employed, which, although it is not
permanently lost, is locked up in the apparatus, and, therefore,
adds considerably to the capital outlay in making the installation.
/ afierwarda attempted to replace the copper by f raiments ofcoke^ and
apparently with complete succeeSyfor the copper coats the/ragmenla
80 thoroughly and so easily that they are able at once to form a
perfect substitute for the granulated copper"
The Elkington Patents. — Elkington's process possesses great
interest in connection with copper refining, as it was the first to
utilise the current from the dynamo. The English patent* dates
from the year 1865, the American t bears a considerably later
date. The following is a transcript of the later British specifi-
cation, omitting preliminary matter : —
" This invention has for its object improvements in the manu-
facture of copper and in separating other metals therefrom.
" For this purpose I smelt the copper ore so far as to obtain
an impure metal therefrom, which I then cast into plates, and by
means of electricity I dissolve these plates and deposit the pure
copper on to other plates. The other metals with which the
* English Patent 2,838, Nov. 3, 1865 ; and 3,120, Oct. 27, 1869.
tU.S.A. Patent 10,131, of Feb. 22, 1870.
Digitized by VjOOQ IC
COPPER. 18^
copper was combined fall for the most part to the bottom of the
Yessel in which I operate. This process in its general outline as
above stated is not new, for it has before been patented and used
by me. (See patent dated November 3rd, 1865, No. 2,838.)
''My present invention consists in improvements in th&
method of conducting this process. I prefer to employ copper
ores which contain sufficient silver materially to injure the
copper if smelted in the ordinary way, and which, consequently,
would usually be submitted to a process for extracting the silver
before they are smelted. In such ores, frequently, the quantity
of silver is not such as to pay for the cost of extraction, but the
process has, nevertheless, been necessary when copper of high
quality is required to prevent injury to the copper. These ores
are particularly suitable for my use, as the silver they contain,
which does not raise their price in the market, is recovered by
me without any additional cost. Ores containing a larger quan-
tity of silver, say from 8 ounces to the ton and upwards, and
which are now always submitted to a process for extracting the
silver before they are smelted, can also be advantageously
worked by my process, as can also ores containing little or no
silver, but in this latter case the advantage of my process over
the ordinary process is mainly in the better quality of copper
which I obtain. I smelt the ore in the usual way so as to obtain
all its metallic contents (except such as may be volatile) in the
form of a regulus, from which stage, by preference, but it is not
essential, I carry the metal on to the state of pimple or blister
copper ; this impure metal I cast into plates, say 24 ins. long, 8
ins. wide, and 1 in. thick. One end of the plate is provided at
the centre with a stout T-shaped head of wrought copper ; it is
placed in the mould in which the plate is cast. Cast-iron moulds
are used. The metal is tapped out of the furnace on to a sand
floor, and is led by channels into the moulds. The plates thus
cast are ready to go to the dissolving-house, which is laid with a
wooden floor, inclined from end to end ^ an inch to the foot.
The boards are grooved on their edges, and small strips or
tongues of wood are inserted into the grooves so that there may
be no open joints, and the surface is thoroughly saturated and
coated with pitch to make it watertight. The surface of the
floor is divided into a nuu)ber of troughs running from end to
end of the building by ledges of wood fixed down upon it; these
are also saturated with pitch. Each trough is of a width to
receive three stoneware jars side by side. The jars are cylindri-
cal, 34 ins. high and 18 ins. wide. There are pathways. between
the troughs for the workmen who attend to the process. Each
trough is filled from end to end with jars. There may be, say,
about 100 jars in each trough, and 12 troughs in the width of
the building. The jars should be of fireclay ware, so that they
may not be injured by the solution which they receive. Each
Digitized by V^jOOQ IC
190 ELECTRIC SMELTING AND REFINING.
has a hole in the bottom, closed by a wooden plug, also a hole in
the side 4 ins. from the bottom, and another hole diametrically
opposite to the first and 4 ins. from the top. The jars are set up
level on the inclined floor with wo3den wedges saturated with
pitch. The jars are connected together from the upper to the
lower end of the room, each jar having a pipe passing out from
it at the hole near thQ top and entering the next jar below at
the hole near the bottom. The connection with the jars are
made with vulcanised india-rubber, and intermediate of the
connections the pipes may be of lead, and about | in. internal
diameter.
" The solution which I employ is water charged with as much
sulphate of copper as it will dissolve. The sulphate of copper of
commerce may be used, or for economy I sometimes use a solu-
tion obtained by boiling the deposit found in the culvert or long
flue by which the smoke from the copper furnace is led to the
high chimney; this will furnish a solution of sulphate of copper
sufficiently pure for the purpose.
''The solution is stored in a tank at the upper end of the
dissolving-room; it is admitted into the uppermost jars, and runs
from jar to jar until those at the lower end of the building are
filled. Clips are put upon the india-rubber connections to stop
the flow through the tubes when tlie jars are full, and so to
maintain the solution at the proper level in the upper jars.
" When the process is in operation two (tic) clips are taken
ofiP, say, once in twenty-four hours, so as to cause the solution to
flow through all the jars and transfer the solution from the
bottom of one jar to the top of the next, so as thoroughly to mix
it, as in working it tends to become weak at the top of the jar.
This means of readily equalising the density of the solution is
of great practical importance. At the lower end of the room is
a tank to receive the solution. When it is thus allowed to run
through it is pumped back into the upper reservoir. The same
tank receives the contents of the jars when they are emptied on
to the floor beneath by the removal of the bottom plug, as here-
inafter described. In the gangways between the floor-troughs a
truck runs to carry the cast copper plates to the jars in which
they are to be dissolved. Six metal plates are suspended in
each jar; they are hung in couples from the horizontal copper
bars, having forks upon them to receive the T-form heads of the
plates. These bars rest at their ends on other bars of wood laid
on the jars so as each to extend across a row of three jars, and
the same bars also support over each jar two other metal cross-
bars to support plates to receive the deposit of copper from the
solution. There are four receiving plates in a jar, two suspended
from each bar; they are interposed between the cast plates.
Conducting strips of sheet copper are laid upon the wooden bars
so as to couple the cast plates of one jar to the receiving plates
Digitized by V^OO^ It!
COPPER. 191
of the next jar, and so throughout the series of, say, 100 jars.
Each metal cross-bar is made to bear on a connecting strip at
one end, and at the other on a wooden block saturated with
pitch. The jars are each provided with a false bottom of wood
to prevent breakage of the jar in case a plate should fall. The
receiving plates may be of wrought copper, but I prefer to em-
ploy in the first instance guttapercha coated with bronze powder.
As soon as a deposition of copper is obtained, the guttapercha is
stripped off and the copper left to receive a further deposit. A
series of, say, 100 jars being thus coupled up into a circuit, I
connect to the terminals of the series one or more electro-
magnetic machines. I prefer to employ the machines manufac-
tured by Messrs. H. Wilde <fe Co., of Manchester. The machines
called by the makers 34-inch machines are those which I use,
and I drive them at 2,500 revolutions per minute. With three
such machines working into a series of 100 jars a deposition of
4 or 5 lbs. of copper in each jar may be obtained in twenty-four
hours without injury to the solution. When the cast plates
become so far dissolved as to be unfit for further use they are
removed; their remains are washed in the lower solution tank
to remove the deposit from their surfaces, and they are melted
and recast. The wrought T-heads may be used an indefinite
number of times, as I protect them from solution by coating
their stems with wax. The receiving plates are allowed to grow
until they attain a convenient weight ; they may either be
melted and cast into cakes, and afterwards rolled in the usual
way of working copper, or the plates as they come from the vats,
may be sent into the market. The solution may be worked for
a very long time, evaporation being supplied by the addition of
water acidulated slightly with sulphuric acid, and ultimately
will become so charged with sulphate of iron as to make it
inconvenient to work it further. If, however, the metal be
advanced to the pimple or blister stage before casting the plates,
it will take but little iron into the solution. The silver or other
metals (excepting the iron) with which the copper of the cast
plates was contaminated, sinks to the bottom of the jars and is
there allowed to accumulate until it reaches the lower side hole;
when this happens the bottom plugs are taken out of all the jars
of the series, and the contents washed out into the floor trough,
which discharges them into the tank at the end of the building ;
here they settle, from time to time the tank is pumped dry,
and the sediment is taken out. There are two such tanks at
the lower end of the room to allow of one being put out of use
before emptying.
"The sediment may be treated in any ordinary and well-
known manner for the recovery of the sUver it contains, and
other metals may be separated from it should it be considered
desirable to do so."
Digitized by VjOOQ IC
192 ELECTRIO SMELTING AND REFINING.
Excepting the form of the decomposing vessels, which is a
little unpractical, and the method of arranging the electrodes,
which is lacking in simplicity, this process is one which exhibits
all the essential features of the modern methodn of copper
refining; and it must be admitted that £lkington well under-
stood how at once to turn to his advantage the discoveries in
electro-technology.
Theory of Electrolytio Copper Beflning. — The process as
it is practically applied consists, then, in the electrolysis of a
copper sulphate solmtion with anodes of crude copper and pure
copper cathodes. And whether the phenomena observable in
carrying out the process be interpreted by the old explanation,
or by the new, it will be found that no great expenditure of
power can theoretically be necessary to carry over the copper
from the positive to the n^ative electrode. According to the
older view the electric current serves to decompose the electro-
lyte ; and it would therefore first cause the transposition shown
in equation (i.) : —
(1.) CUSO4 + H2O = Cu + HjS04 + O.
(-) (+)
But then at the same time there would be occurring an
oxidation and solution of copper at the anode, with a reproduc-
tion of the original electrolyte, and therefore a reversal of the
former equation, thus : —
(ii.) Cu + H2SO4 + O = CUSO4 + HjO.
On calculating the E.M.F. necessary to these reactions from
the thermo-chemical data indicated by the two equations, it will
be seen that the one balances the other, and the resultant is nil.
According to the modem explanation, the electrolytes are
entirely or in part dissociated. In such a case the solution must
contain equivalent Ou'' ions and SO4'' ions. The solution is
nearly saturated, and the osmotic pressure is therefore more
favourable to the separating out of dissolved substances. But
another characteristic of copper refining is that large quantities
of copper are transported from the anode to the cathode. It
is true that the copper of the anode possesses a solution pressure,
even if it be but small ; but equilibrium is maintained without
any further expenditure of energy by the back osmotic pressure
of the cations already existing in the solution. But even a
slight excess pressure, produced outside the bath by the dynamo,
which is in direct metallic connection with the electrodes, serves
at once to set in motion large quantities of copper on their path
from the anode to the cathode, where an escape is, of course,
provided for the electrical charges given up by the copper ions
as they deposit.
Theory and practice alike teach that only a very small electro-
Digitized by V^OOy It!
COPPER. 193
motive force is required to drive copper over from the anode to
the cathode, leaving its impurities behind at the anode, and
that, therefore, the quantity of electricity that it is arranged
shall be carried by the copper ions must be very large in pro-
portion to the E.M.F.
Simple as the theoretical explanations of the process may be,
it is not quite so simple in actual practice. The copper anodes
are not composed of pure copper ; they contain a considerable
proportion of foreign matter which, if left insoluble, is capable
of producing polarisation* For this reason the anodes and the
electrolyte, and therefore the necessary KM.F. also, will alter
from the very beginning of the operation. Kiliani, in the year
1885,* undertook the task of accurately examining and recording
the behaviour of the impurities present in copper under electro-
lysis. With reference to the conditions of work in this method
of refining, he writes : —
S[iliani'8 Experiments in the Behaviour of Impurities
during the Eleotrolytio Beflning of Copper. — *'With a
normal current-density of 20 amperes per square metre [1*8
amperes per square foot], and a solution containing 150 grms. of
copper sulphate and 50 grms. of sulphuric acid per litre [1 56 lbs.
copper salt) and 0*52 lb. acid per gallon], cuproiu oxide in the
anode remains unaltered, because it is a very bad conductor,
and passes at first into the slime or mud left at the electrode ;
secondarily, however, it gradually dissolves in the acid of the
bath, and thus passes into solution at a rate proportional to
the time that it is left in the liquid. The presence of cuprous
oxide in the anode, therefore, causes the bath to become less
acid, but richer in copper.
**Stilphide (or sdenide) of Copper passes as such into the slime,
so long as it is not present in quantities which would be con-
sidered abnormal in black copper, and especially if it exists in
the form of sub-sulphide (Cu^S). Sulphur cannot then be ex-
tracted from the slime by means of carbon bisulphide. It is
only when the anode contains a considerable proportion of
sulphide, as in the case of matte, that this compound is decom-
posed with separation of sulphur.
'* Silver, PlaHmimy cmd Gold remain completely in the form of
powder in the slime, provided that they be not present in very
considerable proportion, and that the solution contain the normal
quantity of copper and add. If, however, the solution should
become neutral, the silver will rapidly pass into solution, and
will then, of course, be deposited at the cathode.
^ Bismuth cmd Bismuth Oxide in part pass directly into the
slime and in part dissolve into the solution, from which they
separate almost completely, in course of time, in the form of a
basic salt
* Berg- und HmenmSnmeches Zeitung, 1885, p. 249.
13
Digitized by LjOOQ IC
194 ELECTRIC 8MELTINO AND REFINING.
"Tin at first dissolves in the bath, but partly deposits again, on
standing, as a basic salt ; but if much of this metal be present
in the anodes, the greater part will be left as basic sulphate on
the anode itself. In the moist condition, this anode slime has a
dirty pale grey colour, but on drying in the air it becomes
white, and increases rapidly in weight, even after long drying
at 100° 0., so that it is almost impossible to weigh it correctly.
Finally, the slime contains, besides sulphuric acid, tin oxide,
which is present almost entirely in the form of the a-stannic
acid, soluble in hydrochloric acid, and only in very small pro-
portion as the P' (meta) stannic acid, which is insoluble in that
liquid. The presence of tin, therefore, causes the bath to become
weaker in respect of copper, without gaining any considerable
proportion of tin in exchange ; and in consequence of the
separation of the basic salt, the liquid becomes slightly, but
quite appreciably, more acid.
"But the presence of the tin in the solution appears to exert a
remarkably favourable influence upon the deposit of copper; for,
whilst a bath consisting of a chemically pure neutral solution
of copper sulphate gave, with the above-named current-density,
a deposit that was exceedingly bad, warty, and brittle, the same
current produced an exceptionally good copper, free from wart-
like protuberances, and very malleable, when the anode con-
tained a notable proportion of tin. This favourable result was
obtained even when no trace of tin was to be found in the electro-
deposited copper ; it was tested for by dissolving the deposit in
nitric acid, evaporating, and re-dissolving the residue in acidu-
lated water, and not the slightest precipitate, or even turbidity,
indicative of the presence of tin, could be observed in the re-
sulting solution. From the observation of this phenomenon,
no doubt, arose the custom, practised by a few electrotypers, of
adding tin salts to their baths. The E.M.F. required for the
electrolysis is also markedly reduced when fairly large propor-
tions of tin are present in the anodes.
*^ Arsenic dissolves in either acid or neutral solutions as
arsenic acid, until the solution is saturated therewith, and then
it remains undissolved in the slime. The arsenious acid, com-
bined with cuprous oxide or other metallic oxides, passes
entirely into the slime if the solution be neutral, because these
compounds are not conductors of electricity. If, however, the
electrolyte be acid, a secondary (purely solvent) action takes
place, and the arsenic slowly dissolves into the bath in the form
of arsenious acid ; but this solution, of course, takes place in a
less degree the more often the action of the acid is stopped by
lifting the anodes from the vat and removii^ the slime from
their surfaces. The influence of the arsenic in the anode copper
is, therefore, to make the bath poorer in respect of copper, but
richer in acid.* The copper deposited from neutral arsenical
Digitized by V^jOO^ It!
COPPER. 195
solutions will contain arsenic, but that from acid solutions will
be free from this impurity unless the bath contain proportion-
ately very little copper.
^^ArUimony goes partly into solution whether the bath be acid
or neutral, and in part remains as basic sulphate at the anode,
whilst a portion of that which dissolves separates out again on
long standing. The antimonial anode slime behaves like that
containing tin, in that it gains in weight on exposure to the
air. Antimony, therefore, makes the electrolyte poorer in
copper. The antimoniates are not decomposed by the current,
and therefore remain at first in the slime ; they are, however,
gradually attacked by the acid of the bath with separation of
antimonic acid, and so tend towards the neutralisation of the
electrolyte ; but, of course, this effect is the least marked when
the slime is the most often removed from the bath. Even when
the solution is saturated with antimony, and basic salt is be-
coming precipitated in the vats, the antimony is not deposited
at the cathode so long as the electrolyte contains approximately
the normal proportions of acid and copper ; at the worst, some
basic salt may in such a case mechanically adhere to the cathode
in the shape of a black muddy deposit containing copper and
antimony. But if the solution should be nearly or quite
neutral, antimony will come down with the copper, and the
deposit will be dun coloured and brittle, and will often be
characterised by long needle-like excrescences, bounded by
rectilineal planes. So, too, if the electrolyte contain too little
copper salt, even the usual proportion of acid will not prevent
the antimony being precipitated with the copper.
^^ Lead J under the action of the current, is attacked even before
the copper, and passes into the slime as insoluble sulphate, only
traces going into solution, without any tendency to deposit at
the cathode. The presence of lead in the anode, therefore, makes
the bath proportionately poorer in copper.
"Iron, Zinc, Nickel, and Cobalt dissolve under the action of the
current more readily than copper, and, therefore, weaken the
solution in respect of the latter metal. But besides this, when
small current-densities are employed (as they are in copper
refining) a simple chemical action takes place between these
metals and the free acids present, which results in a greater
solution of such metals at the anode than is equivalent to the
copper deposited at the cathode; the bath, therefore, loses in
free acid, but contains in exchange a greater proportion of
metallic salts.
**Iron always forms a ferrous compound on dissolving when the
current-density is small, but this is gradually peroxidised to the
ferric state under the influence of the air during the circulation
in the vats j and this reaction also tends to a neutralisation of
acid. Ferric salts are formed at the anode itself only when a
Digitized by V^OOQ IC
196 ELECTRIC SMELTING AND BKFININO.
high current-density of about 1,300 amperes per square metre
[120 amperes per sq. ft.] leads to a separation of oxygen and
free acid at that electrode. Sulphide of iron at the anode forms
only ferric salts. Even if all the copper in the solution be
replaced by iron, excepting about 2 grms. per litre [^ oz. per
gall.], warty excrescences will still be formed.
'* The mud or slime deposited at the anode may contain, after
drying, gold, platinum, silver, silver sulphide, cuprous oxide
and sulphide, basic sulphates of bismuth, tin and antimony,
antimonic acid, arseniate of copper, metallic arseniates and anti-
moniates, lead sulphate, and slag constituents, in which may be
iron, lime, magnesia, and silica ; and with these there will also
be a certain amount of metallic copper in pulverulent form.
The gradual solution of the anodes does not always so take place
that the ianermost portions only are attacked when the outer
surfaces have been completely dissolved ; on the contrary, it
happens much more frequently that the solvent action penetrates
far into the interior of the anode while there is yet much copper
at the surface. Brittle black copper plates, for example, con-
taining 96 per cent, of copper, 2'5 mm. [^ in.] thick, were
submitted to electrolytic solution, and even after ten days in
the bath they had become so soft throughout that they could be
rolled together like cardboard without breaking. Anode plates
cast direct from copper pyrites always exhibited a marked
bulging towards the cathode after a few days in the electrolytic
tanks. The electrolyte itself usually becomes weaker in respect
of acid and copper, whilst taking up iron, zinc, nickel, cobalt,
manganese, tin, arsenic, antimony, and bismuth, and it then
shows a higher total proportion of metallic salts.
"The last-named disadvantage is, however, to some extent,
compensated by a by-reaction; for it always happens, in the
treatment of comparatively pure samples of crude copper, that
the solution becomes gradually richer in copper, without sufficient
evaporation taking place to account for the phenomenon. It is
a well-known fact that copper is able to exert a slight reducing
influence on acid solutions of copper sulphate with the formation
of a little cuprous salt, which then, under the action of the air,
becomes reconverted into cupric sulphate. This peculiarity was
studied by H. Boessler,* and formed the groundwork of the
older sulphuric acid process of treating copper. This by-reaction
leads to a slight but continuous solution of copper, which is
always the more marked as the current-density is reduced, and
the circulation of liquid is made so much more rapid, that
the solution is brought into better contact with the air. The
solvent action thus caused is most marked in the neighbour-
hood of the upper surface of the bath, and may be so great
that a thin cathode plate projecting above the electrolyte will
* Dingier^ 8 polytech, Joum., vol. ocxlii., 1881, p. 286.
Digitized by V^jOOQ IC
COPPER. 197
be completely cut through at this level in the course of eight
days. This by-reaction explains the fact that the loss of weight
at the anode is greater, and the gain at the cathode is somewhat
less, than they should be respectively if they corresponded to
the current-strength employed.
" It is Hverefwe very important that the percentage of acid in the
rbath should be determined from time to time, and tJuU any de-
Jiciency slioiUd be made good. And it is equally necessary that
the proportion of copper shall not be allowed to fall too low.
The most favourable current-density is 20 to 30 amperes per
isq. metre [1*8 to 2*8 per sq. ft.].
" The gradual neutralisation of the solution, produced by the
operation of the various reactions above alluded to, has the most
detrimental influence on the whole course of the electrolysis.
In the first place, the conductivity of the electrolyte is greatly
diminished, and the difference of potential required between the
electrodes, under otherwise normal conditions (with the elec-
trodes 5 cm. [2 ins.] apart), may have to be raised from 0*1 to
0*25 volt, solely on account of this neutralisation. Then the
impurities present in the bath are liable to pass into the cathode
deposit as already explained, and to make it brittle and useless.
But apart from the introduction of impurities, the deposit
obtained, even from a chemically pure solution, is exceedingly
unsatisfactory if the liquid be neutral ; and it may in fact be so
brittle that it can be crushed to powder in a mortar. The cause
of this trouble is the formation of cuprous oxide. When the
current-density is small, the current does not decompose the
copper sulphate completely into metallic copper and SO4, but it
deposits a certain proportion of cuprous oxide as well, and the
amount of the oxide diminishes as the current-density is in-
creased, until at a certain limiting point pure copper is deposited
alone. In acid solutions this cuprous oxide becomes decomposed
by a secondary, or chemical, action, whereas in neutral liquids it
remains attached to the cathode.
"A good circulation of the electrolyte is also an essential,
since, otherwise, the upper portions of the bath will be more or
less denuded of copper, and impurities will in consequence be
deposited on the corresponding portions of the cathode. The
influence of the circulation on the potential in normal solutions
cannot well be measured; but the E.M.F. required increases
markedly with any addition of impurities. This is shown by
experiments, the results of which are quoted in the following
table. In these, cathodes of pure copper were used, with a
distance of 5 cm. [2 ins.] between the electrodes, and a current
of 20 amperes per sq. metre [1*86 amperes per sq. ft.} The
black copper anode used contained per cent. — 96*6 of copper,
0*403 of silver, 0*011 of gold, 1*23 of arsenic, 1 of iron, and
0*54 of sulphur.
Digitized by LjOOQ IC
198
ELECTRIC SMELTING AND REFINING.
Table showing Effbct of Impubities on the E.M.F. required
FOR THE ElROTRO-DEPOSITION OF Ck>PPER.
1,000 parts of solution (by measure^
contained.
Anode.
E.M.F. in Tolta.
With
circula-
tion of
solution.
Without
drcula-
tlonof
SOlutiOD.
Weight.
160 parts copper sulphate, )
50 „ sulphuric acid, j
Pure copper, . .
Black copper, . .
Copper matte, .
0-095
0120
0-400
0-096
0120
0-400
160 parts copper sulphate,
Pure copper, . .
Black copper, . .
Copper matte, .
0-240
0-276
0-532
0-243
0-278
0-535
0-75
076
1-00
7-96 parts copper sulphate, )
158-2 „ ferrous sulphate, >
50*0 „ sulphuric acid, )
Pure copper, . .
Black copper, . .
Copper matte, .
0-22
0-25
0-50
7-96 parts copper sulphate, J
168-2 „ ferrous sulphate, J
Pure copper, . .
Black copper, . .
Copper matte, .
030
0-35
0-76
1-10
1-15
1-30
<* The absence of circulation in the solution, however, not only
necessitates an increase in E.M.E., but it exerts a most serious
influence upon the physical and chemical properties of the
copper deposit ; for this is always purer, more finely crystalline,
and more malleable in proportion as the liquid is well stirred^
even when perfectly pure solutions are used, and otherwise
normal conditions observed.
" In order to calculate the expenditure of energy necessary for
the treatment of a given raw material, a laboratory experiment
must be made in which the difference of potential between the
two electrodes is measured, when they are separated by a space
equal to that which will divide them in practice, and when they
are subjected to the same current-density that has been proved
the most favourable on the large scale. If, then, for example,
potential difference at the poles of the dynamo be 15 volts, and
that required for each pair of electrodes be 0*25 volt (neglecting
for the moment the external resistance of the vat connections),
it should be possible to couple at most (15 -^ 0*25 » ) 60 pairs
in series ; but this number is never attainable in practice, and
on an average 40 baths so arranged would be the more probable
Digitized by V^OO^ It!
COPPER. 199
practice. If now the dynamo afforded a current of 240 amperes
at the above voltage, corresponding to a deposit of 283*61 grms.
[0*625 lb.] of copper per hour, there would then be obtained in all
the 40 baths arranged in series a total of 11,344 grms. [25 lbs.] of
copper in one hour, or 272*26 kgs. [600 lbs. in twenty-four hours.
The power required to accomplish this is (240 x 15) -f 736 = 4*9
[German] H.P. for the dynamo, or about 6 [German] H.P. for
the steam engine ; [or (240 x 15) -^ 746 = 4*8 H.P. and 5*9 H.P.
respectively, calculated according to the British unit]. It should
be remembered that a plant of this capacity will require a
superficial area of 80 sq. metres [860 sq. ft.], and that with the
normal current-density of 20 amperes ])er sq. metre [1*86 amperes
per sq. ft.] five months will be required to produce a copper
plate 1 cm. thick,*' [or 6^ months to deposit a plate ^ in. in
thickness].
Wohlwill's Experiments : Formation of Copper - Mud
at Anode. — Modern theory is not in accord with all Kiliani's
explanations, and the views of so competent an electro-metal-
lurgist as Dr. E. Wohlwill,* for many years the manager of the
electrolytic copper, silver, and gold refinery of the Norddeutache
Affinerie at Hamburg, will therefore be specially welcome.
*'In Kiliani's paper, published in 1885, which is commonly
regarded as the standard explanation of the chemical and electro-
chemical processes of the electrolytic copper refinery, sufficient
attention was not paid to the presence of considerable quantities
of finely-divided metallic copper in the anode-mud. I [ Wohlwill]
have, therefore, studied the phenomena in minute detail, using,
in the first place, anodes of pure rolled electrolytic copper. £ven
with these anodes there is always a separation of finely-divided
copper dust on their surface when the current is flowing, but its
colour is pure red because it is not mixed with any foreign
matter. It adheres but lightly to the surface of polished plates,
and always precipitates in part to the bottom of the containing
vessel; the remainder is readily detached by rubbing with a
camel's-hair pencil or with the aid of a [chemical] washing-bottle.
It is difficult, however, to obtain an accurate determination of
the quantity of the loose deposit formed under varying conditions,
because, in the presence of air, it is very soluble in the acid
solutions which are, of preference, used for the depositing pro-
cess. An approximate estimate of the copper lost as mud may
be formed by deducting the amount lepresenting the increase in
weight of the cathode at the end of the operation from that
representing the loss in weight of the anode in the same period.
From estimations made in this way it appears that : —
(1) The quantity of the anode-mud formed depends mainly on
the current^ensity — ^the smaller the current-density the greater
is the quantity of waste deposit in a given time ;
* From a private communication*
Digitized by LjOOQ IC
200 ELECTRIC SMELTING AMD REFINING.
(2) With equal current-densities the amount of the waste
deposit increases with the acidity of the electrolyte ;
(3^ Given otherwise similar conditions of current-density and
acidity, the quantity of the waste deposit, measured by the
relatively greater difference in weight of the anode, is propor-
tionately smaller the longer the interrupted action of the current
has continued.
** Since not the slightest separation of finely-divided particles
of metal is to be observed when pure copper is dissolved by
simple chemical solution— as, for example, when immersed in a
solution of ferric sulphate — it is to be assumed that the formation
of the copper-mud is entirely due to electro-chemical action.
" The different behaviour of the same metal when exposed to
the solvent action of the current oould be explained by the for-
mation of a small proportion of cuprous sulphate along with the
greatly preponderating mass of cupric sulphate at the surface of
a copper anode immersed in a sulphuric acid solution, the cuprous
salt decomposing into cupric sulphate and metallic copper in
immediate, or almost immediate, contact with the anode. Ac-
cording to this explanation, the copper-mud is a product, not of
decomposition of the anode, but of the cuprous ions passing into
the solution. The self-same action that occurs without any
assistance from an electric current when copper dissolves in a hot
solution of copper sulphate and deposits again on cooling, is,
therefore, here repeated in a very narrow space at the anode.
(This action, which has been repeatedly referred to of late years,
was, to the best of the writer's knowledge, first described by
Rossler.*) Theeffectof usingalowercurrent-density,and therefoi-e
causing a greater sub-division of the same quantity of electricity,
is, that the proportion of monovalent ions to divalent ions is
increased, as is also the quantity of waste copper produced from
them ; and this observation corresponds exactly with the mental
picture that is formed of the process of ionisation.
''In the writer's opinion there must also be further efifects
resulting from the deposited copper-mud which is in contact with
copper that is as yet unattacked, behaving as an insoluble or, at
least, as a less attackable or more negative metal. If this be so
the mud must not only remain uninfluenced by the solvent
action of the current, but must protect the part of the anode
lying immediately beneath it from contact with the electrolyte,
and, therefore, from the influence of the current. The solvent
action would thus be exerted on the adjacent unprotected parts
of the anode. In this way the available unattackable surface
will be to some extent reduced owing to the separation of the
muddy deposit, and in consequence the current-density must
increase and the number of monovalent ions produced in a given
pet iod of time must be proportionately diminished. It is thus
• Dingier^ s poly tech. J<mm,, 1881, vol. ccxlii., p. 286.
Digitized by
Google
COPPER. 201
«asy to account for the observation that, within certain limits,
the increased loss of the anode per hour becomes less as the time
of the electrolytic action is extended.* But such an increase in
current-density unaccompanied by any alteration in the yolume
of current must lead to the action being relatively intensified on
the smaller surfaces exposed, and consequently to the formation
of abrupt variations in the level of the surface, or, in other words,
to the roughening of the anode. This roughening effect is
invariably to be observed in the case of copper anodes which have
been exposed for some time to electrolytic action ; and it always
becomes more marked as the duration of electrolysis is extended.
Chemical solution, on the other hand, unaccompanied by the
formation of such mud-like deposits, does not impair the even-
ness of surfaces which are initially smooth, but, on the contrary,
tends to round off any projecting comers that there may be.
" The surface of copper anodes may remain smooth even when
undergoing electro-chemical solution, provided that the current-
density applied is so high that no considerable number of cuprous
ions are formed, and the amount of copper-mud deposited is there-
Fig. 110.— Diagram illustrating irregular solution of copper anodes.
fore inappreciable. In the case of surfaces which have become
roughened by use as anodes, a considerable quantity of the copper-
mud clings to them by adhesion during the time of deposition,
and its removal mechanically is by no means complete. PJates
of this kind, therefore, being subject to an apparently variable
loss, are not suited to demonstrate the correctness of the laws
above referred to. The covering of copper -mud imparts a dark
or bright red colour to the plates. A disintegration of the
anode, leading to the breaking away of small fragments, may
also result from the covering of the anode with a protective film
of anode -mud if, in consequence of the action, the shielded
portions are surrounded with less perfectly protected depressions.
Under these circumstances, as shown in Fig. 110, a surface, a,
protected by the deposit, may be completely undermined by the
solvent action that is being brought to bear on the sides of the
cavities, b and c, and in time fragments, large or small, as the
case may be, will become detached.
*' The entire absence of such a precipitate when copper is
electro-chemically dissolved in hydrochloric acid and other
chlorides serves to confirm the views above explained as to the
origin of the copper anode-mud. The cuprous ions formed at the
* The same result must ensue if a portion of the protecting copper is also
dissolved when the area of the uncovered part of the metal is reduced.
Digitized by ^^jOOQ IC
202 ELECTRIC SMELTING AND BEFINING.
anode are able to dissolve in the chloride solution in any propor-
tion, so that there is no discharge of ions at the anode and no
separation of copper as in the case of the sulphate bath. Further,
the quantity of copper dissolving at the anode is independent of
the current-density, and exactly corresponds in all cases to the
equivalent of monovalent copper. If the concentration of the
chloride in the electrolyte is low there will be a separation of
white cuprous chloride at the anode. The same phenomenon is
observed if sodium chloride or hydrochloric acid be added to the
sulphate bath, since the added chloride at once reacts with what-
ever cuprous sulphate is formed at the anode and produces the
insoluble cuprous chloride. In this case, as in the last, there is
no formation of metallic mud.
" The separation of the red anode deposit is a characteristic
indication of the purity of the copper. Even a very small pro-
portion of impurity can be recognised by the blackening of the
deposit and of the anode itself Hence, all copper produced
otherwise than electrolytically will become black when used as
an anode, as will also electrolytic copper in the deposition of
which sufficient care has not been taken to avoid the co-precipi-
tation of foreign substances, and especially of arsenic and anti-
mony. It is, therefore, easy to obtain a control for the proper
conduct of the electrolytic process by occasionally transferring
a test plate from the cathode to the anode side of the vat.
" The anode-mud of black copper only differs from the black
anode deposit of imperfectly purified fine copper in the quantity
of the insoluble constituents admixed with it. The proportion
of metallio-cof>per dust contained in the mud depends upon the
same laws that govern the deposit on the pure copper anodes,
and therefore always becomes greater when the current-density
used for the decomposition is reduced. With lower current-
densities, such, for example, as are used on a large scale in the
Norddeutsche Affinerie at Hamburg — viz., 40 to 50 amperes per
sq. metre [3*7 to 4*6 amp. per sq. ft.l — the quantity of copper in
the anode-mud is very considerable.*
'*The irregular corrosion of the surface and the breaking
away of fragments and particles of various sizes is much more
marked with black copper anodes than with those of pure
copper, since the protective action of the copper-mud and its
attendant consequences are increased by the addition of the
insoluble or difficultly-soluble constituents of the black copper.
"The presence of finely-divided metallic copper in the anode-
mud furnishes the best explanation of the gradual increase of
copper sulphate and decrease of free acid observed in the baths
• It appears that Kiliani must have regarded what was actiially copper
as being cuprous oxide. This oxide can only occur in anode-mud to the
extent that it already exists in the black copper, and has escaped
decomposition by acid.
Digitized by LjOOy It!
COPPER. 203
used in electrolytic refining. The copper in this form is readily
dissolved as copper sulphate in the presence of sulphuric acid
and air; hence, in those refineries which employ a current of
air to agitate the bath and so overcome irregularities of con-
centration, there is a notable consumption of free acid and
formation of sulphate. There is no doubt that the few cuprous
ions which exist in a cold solution, and the chemical action of
the dilute acid on the cathode copper, which is commonly
regarded as the chief cause of the changes in the electrolyte,
only exert a small influence in this respect as compared with
the finely-divided copper in the anode-mud. The secondary
reactions above referred to do not suffice to explain the marked
increase that is commonly observed in electrolytic retineries in
the amount of copper sulphate in the bath, even when the anode
material contains a considerable proportion of such impurities
as teud by their solution to deposit copper od the cathode,
and so to lessen the proportion of copper sulphate in solution.
Although there was no lack of causes for such an impoverish-
ment of the solution in the works of the Norddeatsche Affineris,
it regularly happened tliroughout several years of work that
about 2 per cent, or more of the anode copper was recovered
in the form of sulphate instead of as metallic copper. Without
doubt the continual use of compressed air to mix the solutions
contributes to the formation of a large proportion of copper
sulphate in the baths. The quantity of this salt ultimately
produced would not, however, be less if mechanical stirrers or
gases free from oxygen were employed to mix the solution, be-
cause that portion of the finely-divided copper which ordinarily
enters into the bath as sulphate would in the end have to be
recovered in the form of sulphate, by the subsequent treatment
of the anode-mud outside the bath.
<< The Behayioiir of Chlorine in the Eleotrolyte. — It
should be remarked that chlorine in combination with copper
is a never-failing constituent of the anode-mud produced in the
industrial refining of copper by electrolytic means. After the
explanations which have just been given it is scarcely necessary
to point out that the formation of cuprous as well as cupric ions
at the anode is a pheuomenon that has to be reckoned with.
The production and separation of cuprous chloride in immediate
contact with the anode is a necessary consequence of the forma-
tion of cuprous ions wherever water containing chlorides is
used for the preparation of the electrolyte. It may readily be
observed that during the prolonged electrolysis of solutions of
copper sulphate in river water, the proportion of chlorine in the
water graaually diminishes to almost nil, provided that water
free from chlorine is used to make good the losses by evapora-
tion. Hence, a very extensive precipitation of the chlorine
originally present follows this formation of cuprous ions. The
Digitized by VorOO^ It!
204 ELECTRIC SMELTING AND REFINING.
presence of a relatively large proportion of chlorine in the water
used is immediately recognised by the white deposit on the
anode; but when the quantity is small the colour of the cuprous
chloride that is always formed is liable to be masked by the
impurities in the copper, especially as in the presence of air
the sub-chloride tends to be converted rapidly into a dirty green
coloured oxy chloride. It need scarcely be remarked that, like
finely-divided metallic copper, the sub-chloride in the mad,
decomposed gradually under the combined influence of sulphuric
acid and air, contributes to the increase of copper sulphate in
the solution.
** The cathode, in a smaller degree, plays its part in the
elimination of the chlorine from the water, for the copper
deposited from chloride solutions always contains chlorine, and
this happens also, especially at low current-densities, even when
river water containing a comparatively small proportion of
chlorine is used. The very variable quantity of chlorides in
the waters of the Elbe has afforded the writer many oppor-
tunities of observing this phenomenon. Anodes and cathodes
may be so arranged in the baths that all the rows of anode-
plates are placed between cathode-plates, and the backs of the
outermost cathodes are therefore facing the walls of the tanks ;
the current-density on these surfaces will then be less than it
is elsewhere, and it will be found that, when much chlorine is
present in the water used, even the reverse sides of the end
cathode- plates will be covered with white cuprous chloride,
which rapidly becomes green on ex])Osure to air. Much of
this deposit clings so persistently to the copper that it can only
be entirely removed by hanging the plates as anodes, or, in
other words, by partial re-solution. In the years 1892 and 1893
the amount of chlorine in the Elbe at Mansfeld increased to an
abnormal extent, amounting on September 17, 1892, to 54*33
grms., and on January 25 to 69*5 grms., corresponding respec-
tively to 89*5 and 114*5 grms. of sodium chloride per litre.
The difficulties thus introduced increased proportionately, and
in order to obtain a good red deposit of copper it became
necessary to reverse the connections of the anodes and cathodes
so that both surfaces of the latter should face the anode-plates,
and be subjected to an uniform current-density.
"A smaller quantity of chlorine in the water may be recog-
nised in a characteristic way. It has been customary with the
writer for many years, in addition to the ordinary cathode-plates,
to hang lead strips in the baths as they are brought into use ;
and to remove these strips weekly in order that tiie deposited
copper may be stripped off and examined, and a control be so
kept over the properties of the metallic product of electrolysis.
It is found that the deposited copper, which is always brittle at
first, commonly breaking at the first attempt to bend it even
Digitized by V^jOO^ It!
COPPER. 20&
with great care, improves in dnctility with the lapse of time,
until after a few weeks the plate produced from the same bath
will withstand bending twenty times backwards and forwards
without breaking.
"This decrease in the brittleness of the deposit as the working
period is prolonged is not consistent with the view that tho
cause of the brittleness is the presence in the copper of antimony
or other impurities of that nature, since these impurities are
absent from the newly-prepared solutions and increase with the
age of the bath. The observation that both the brittleness of
the deposit and the proportion of chlorine in the bath steadily
decrease as the time of action is extended, points to the brittle-
ness of the copper deposit being due to the presence of chlorine
in it. The presence of chlorine is readily shown with the aid of
a silver solution, if the copper be dissolved in nitric acid and
the solution evaporated to dryness with an excess of sulphuric
acid. On hanging a piece of brittle copper as an anode in the
bath, the copper free from chlorine dissolves before that which
is not so, and the presence of chlorine in the anode-mud and
waste products is easily recognised.
'* It is clear that this cause of the brittleness of copper, a
cause which the writer believes has not hitherto been noticed,
is worthy of the attention of the electrotypist and electroplater,
who should examine carefully into the proportion of chlorine
contained in the water that they employ.
"On the other hand, it will be understood, after a study of
certain other facts and considerations, that it is not advisable to
use water that is entirely free from chlorine either for electro-
plating and the like, or in the industrial extraction of copper.
** Ocourrenoe of Crystalline Growths on Cathodes. — In
course of the work of the Norddeutsche Ajffmerie the manufacture
of electrolytic copper has on many occasions been delayed and
rendered difficult by the periodical formation of crystals of the
most diverse sizes, offcen as much as 10 cm. [4 ins. J long, on all
parts of the deposited copper. These crystals, which were for
the most part of columnar appearance, proved to be aggregates
of octahedra of varied grouping.''^ Although both the crystals
and the ground mass from which they appeared to spring con*
sisted of pure electrolytic copper, practical considerations
required that the growths should be removed mechanically, and
this led to the expenditure of no small sum in wages. Occasion-
ally, when the epidemic had spread to 600 baths, night work
be^me necessary for several weeks consecutively, in order that
contracts might be completed in the required time.
" It was soon observed that the composition of the Elbe water
* Descriptions of these crystal masses, which are of great interest to
the mineralogist, have been published by Professor Miigge, formerly of
Hamburg, now of Konigsberg.
Digitized by
Google
206 ELECTRIC SMELTING AND REFINING.
played an important part in connection with this phenomenon.
It was found that the crystalline growths vanished almost com-
pletely when (usually in July and August) the proportion of
chlorine in tlie water reached its maximum, and all the anodes
appeared to he covered with a white deposit; whilst, on the
other hand, the trouble was regularly at its worst when
(generally in the early months of the year) the chlorine per-
centage was at its lowest. It appeared, therefore, that the
unknown substance causing these growths in the latter case
reached to the cathodes, whilst, in the former, it was intercepted
either at the anode or in the space between the electrodes.
The views above expressed as to the formation of copper anode-
mud and its prevention by the agency of chlorides present in
the bath, led to the assumption that it must be the cuprous
ions which, having escaped discharge at the anode owing to the
absence of chlorine ions in the electrolyte, produced a separation
of very finely-divided copper at the cathode. The building up
of relatively large growths of copper deposit on these particles
could then be readily understood. The writer has been guided
by this hypothesis in conducting experiments aiming at the
prevention of these crystalline growths. In the first place, at a
time when there was but little chlorine in the Elbe water,
chloride of sodium or magnesium was added to the baths in
sufficient quantity. The action was unmistakable, but it was
found that at best only a diminution of the evil could thus be
effected. The addition of ferric sulphate, which is incompatible
with the existence of cuprous ions in the solution, appeared to
act immediately, but the quantity of oxidisable material in the
baths was so great that a quantity of the salt sufficient for any
length of time could not be employed without disadvantages of
another kind. By means of small scale experiments it was
found that the interposition of suitable diaphragms between the
electrodes afforded a solution of the problem ; and the growths
were entirely absent, even under the most unfavourable con-
ditions of the Elbe water, when the anodes were enclosed in a
wooden casing.
'* A more convenient and satisfactory means of attaining this
end resulted from an observation that there was no growth in
baths in which the electrolytic process was regularly interrupted
day by day, in order that the quantity of deposit might be con-
trolled for experimental purposes by weighing, whilst it often
continued without intermission in all the other baths for weeks
at a time. The introduction of daily interruptions of work for
half an hour throughout the whole plant is not found entirely
to prevent the crystalline growths upon the cathodes, but it
restricts them so far that it is no longer necessary to make
provision for their removal by mechanical means.
" A completely satisfactory explanation had yet to be found,
Digitized by V^jOO^ It!
COPPER. 207
both of the protective effects of an interruption of work and
of the phenomena that have heen described above, but it can
scarcely be doubted that the formation of cuprous ions at the
anode is the actual underlying cause.
*^ The formation of those hair-like growths on copper deposits
which are a source of trouble to the electrotyper is, no doubt,
an allied phenomenon ; and it may therefore be concluded that
it is generally due to a deficiency of chlorine in the water, and
may be obviated by the addition of a little sodium chloride to
the bath."
Fbrster's and SeidePs Experiments. — Forster and Seidel,*
in the course of experiments on the reactions in the copper
voltmeter, have made certain observations bearing on the
question of the existence of a cuprous sulphate, and these have
served to confirm the presence of cuprous ions in sulphate
solutions. From their experiments it would appear that cuprous
sulphate may be formed at the cathode, because, under certain
circumstances, the cupric ions may not be completely discharged
there, and so a portion of the copper is returned into the
solution in the form of cuprous ions. The authors explain the
conditions of this phenomenon as follows : —
1. At current-densities of less than 0*01 ampere per sq. dm.
[0*09 amp. per sq. ft.] the action of the current at the cathode in
concentrated solutions of copper sulphate at ordinary tempera-
tures results solely in the production of cuprous ions. But as
the current-density is increased more cupric ions are completely
discharged, and, in proportion, fewer cuprous ions are produced,
although even at high current-densities the latter phenomenon
is not completely absent.
2. The tendency of the cupric ions in a sulphate solution to
pass into cuprous ions increases very markedly with the tempera-
ture, so that at 100** C, even at current-densities of 0-3 ampere
per sq. dm. [2*8 amp. per sq. ft.], when the solution is con-
centrated, the current forms cuprous ions almost exclusively at
the cathode.
3. The formation of cuprous ions in copper sulphate solutions
may also take place in consequence of the reaction of metallic
copper with cupric ions present in the solution, exactly as in the
case of cupric chloride solutions, thus f —
+ + +
Cu + Cu = 2Chi.
This reaction continues until the cupric sulphate solution is
saturated with cuprous sulphate. It is difficult to decide off-
• ZeiUehrift/ar Eleklrochemie, 1897, vol iii., p. 479.
t ThiB fact was observed many years affo by Jacobi (^. Wiedemann,
Lehre von der EUktr,, ii., p. 510), but the discovery was without inflnenoe
on later work. Cf, also Note on p. 200.
Digitized by VjOOQ IC
208 ELECTRIC SMELTING AND REFINING.
hand hovtr far the formation of cuprons sulphate during the
electrolysis of copper sulphate solutions is dependent on this
process of solution, or, indeed, whether it is a purely electrolytic
phenomenon; but the author considers the latter to be the
simplest and most probable explanation.
4. Under otherwise similar conditions, an increase in the
concentration of the cupric ions, in a solution of copper
sulphate, is accompanied by an increase in the number of
cuprous ions.
5. If the solution is neutral, the cuprous sulphate formed
undergoes hydrolysis (according to the equation :
+ - - + - -
2Cu + SO4 + HjO = CujO 4- 2H + SO4)
as soon as its concentration has exceeded a certain limiting
yalue. In consequence of this, cuprous oxide is frequently
deposited on the cathode in the form of sparkling crystals,
whilst free sulphuric acid is formed in the solution.
6. If the solution is sufficiently acid no hydrolysis ensues,
and there is a much larger number of cuprous ions in the
solution than when it is neutral. But there is a limit here
also, for when the ratio of concentration of cuprous to cupric
ions has exceeded a certain value, the former pass back again
into cupric ions, and metallic copper is deposited thus —
+ + +
2Cu = Cu + Cu.
Hence, referring back to paragraph 3, it will be seen that we
are here dealing with a reversible reaction —
Cu + Cu ;^ 2Cu.
7. From the foregoing it is evident, therefore, that even when
the current is producing only cuprous ions, metallic copper is
formed at the cathode. This copper may be considered as the
product of a secondary reaction, and is precipitated in the form
of small isolated crystals, rather than that of the even, uniform
deposits of electrolytic copper obtained in the usual way from
acid solutions.
8. Ouprous ions aniving at the anode receive positive charges,
and there become converted into cupric ions, so that the current
may be doing other work at the anode in addition to that of
ionising the anode copper.
The observations and explanations given by Forster aad
Seidel are, in all essential points, in agreement with Wohl-
will's experience in industrial work.
Von HiibPs ExperimentB. — Yon Hiibl's'*^ work, although
first published in 1886, may be referred to after that of
*MiUh. des k, u, k, milUcirgeogr, InstU,, 1886, voL vi, p. 51.
Digitized by V^jOOQ IC
COPPER.
209
Wohlwill, Forster, and Seidel, because his observations are
better explained in the light of the work done by the other
three anthors than in his original paper. It may be learned
from an account given by Forster that von Hiibl studied the
effect of current-density on the strength of cathode deposits,
and thus ascertained that both the strength and the hardness
of electrolytic copper increased with the cuiTent-density inde-
pendently of the concentration of the bath, and that with 20 per
cent, solutions of copper sulphate it reached its highest limit at
the density of 2*2 to 3 amperes per sq. dm. [20 to 28 amps, per
sq. ft.], which may be practically admissible. The elastic limit
was highest at 1 to 1*5 amperes per sq. dm. [9*3 to 14*4 amps,
per sq. ft.] in 20 per cent, solutions, but was lower in more
dilute solutions and was higher at 0*85 ampere per sq. dm.
[8 amps, per sq. ft.] than at 1*5 amperes per sq. dm. [14*4 amps,
per sq. ft.]. The toughness was at its greatest at current-
densities of under 0*6 ampere per sq. dm. [5*6 amps, per sq. ft.]
for 20 per cent, solutions of copper sulphate.
BfRsot of Temperature of Bath on Properties of Copper. —
Forster and Seidel* have shown that tiie temperature of the
electrolyte is a further and very important factor in determining
the properties of the copper electrolytically separated from acid
solutions. It appears that copper deposited at 40" C. is especially
fine and uniformly crystalline, and that it possesses great duc-
tility. Forster gives the following numbers, determined by
Hartig, in comparison with the results of similar tests with
wire made from Mansfeld electrolytic copper : —
Temperature
of
Elecstrolyte.
Mean
P.D. of Rath.
Tenacity.
DuctUity
per cent.).
20'
60''
Wire from
copper.
0-32 volt.
0*25 „
0-20 „
electrolytic \
2 16 km.
IS ;:
283 „
912
26 00
13*5
31*0
These figures confirm the results of a long series of bending
tests of copper which had been deposited electrolytically at from
35* to 40"" C, tending to show that an extraordinarily tough
copper is obtained at this temperature. They show, also, that
a further rise in temperature up to 60^ C. results in a reduction
in the ductility of the metal produced, and that the copper
deposited at the higher temperature is stronger than that
*ZeUschri/t/Ur MXektrochemU, 1809, vol. v., p. 608.
14
14 T
Digitized by V^jOOQ IC
210 ELECTRIC SMELTING AND REFINING.
obtained at about 20** C. Nevertheless, the mechanical pro-
perties of electrolytic copper are always distinctly inferior to
those of pure copper which has been mechanically worked.
They approach nearer to it, however, as already shown, when
the electrolyte is maintained at 35** to 40** C, and this tem-
perature must, therefore, be considered as the most suitable for
the production of electrolytic copper with the best mechanical
properties obtainable.
Use of Aloohol in the Copper Solutions. — Finally, refer-
ence must be made to OetteFs experiments, in which the evils
attendant on the formation of cuprous ions in the copper volta-
meter are mitigated by an addition of alcohol to the copper
sulphate solution. Oettel and Forster attribute this action
to a diminution in the electrolytic dissociation of the copper
sulphate, caused by the addition of alcohol. Given that there
are, under these circumstances, fewer free cupric ions, it naturally
follows that the conditions are less favourable to the formation
of cuprous ions.
Development of the Eleotrolsrtio Copper BefLning In-
dustry.— [Elkington, in England, was the pioneer of the
electrolytic copper refining industry, and the works of Messrs.
Elkington, Mason k Co., at Pembrey, in South Wales, seem,
for some time after the 1865 patent was granted (see above,
p. 188), to have been the only place at which the process was
in practical use. They were, however, soon followed by other
works, among which may be named those of Messrs. Vivian k Co.,
at Swansea; Messrs. T. Bolton <b Son, at Oakamoor; Messrs.
Williams Foster, at Swansea; W. A. Hill's, at Chester, and the
English Electro-Metallurgical Co., at Hunslet, as well as the
Elliott's Metal Co., of Selby Oak, near Birmingham, who also
acquired the Pembrey works of Messrs. Elkington, Mason k Co.
In the United States, the Phoenixville Works were refining on
a small scale in 1880, and shortly afterwards Balbacb took up
the work ; but it was not until about ten years later that the
enormous progress was made which has resulted in the colossal
works of the Earitan Coppe'r Co., at Perth Amboy, New Jersey;
the Anaconda Mining Co., at Anaconda; the Baltimore Smeltiug
and Rolling Co., at Baltimore ; the American Smelting and
Refining (5). (Guggenheim Works), at Perth Amboy; the
Boston and Montcuia Consolidated Copper and Silver Mining
Co., at Great Falls, Montana, and the Nichols Chemical Co.,
Laurel Hill, New York, and others.]
The electro-chemical refining of copper in Germany first
gained a foothold in the beginning of the seventh decade of
the nineteenth century in some of the most important copper
works, and from this time developed rapidly. At the outset
the Norddeutsche Affinerie and the Mansfeld Copper Mines
commenced work with experimental plants, and they were soon
Digitized by V^OO^ It!
COPPER. 211
followed by the Communion Hiittenwerk, at Oker; the AUenauer
Kupjerhiitte^ in the Upper Hartz Mountains; the Stadtberger
HUUej in Niedermarsberg ; and C. Schreiber, at Burbach, Siegen;
and, later, by Borchers Bros., in Goslar, in the Hartz, and the
Elmore Melall-AktiengeselUohaft, at Schladern, on the Sieg.
The last-named company produced copper articles, and especially
hollow wares, such as tubes, direct from the crude copper. The
firm of Siemens & Halske, of Berlin, has also done great service
in the development of the industry of electrolytic copper refining,
first in Germany, and then, after gaining the necessary experi-
ence there, in other countries. Even in the newest works there
will be found many of the special characteristics of the original
designs of this firm.
General Arrangement of Plant. — The relatively low cur-
rent-density used in copper-electrolysis necessitates the placing
of many rows of electrodes in each bath, in order to economise
space and to ensure ready access to the apparatus. It was only
natural that, during the early years of the development of the
industry, there should be many mistakes in the choice of
dynamos, in the connections, and in regard to the relation
between the .pressure and the volume of the current. It is,
however, unnecessary in these days to enter into further par-
ticulars of these matters. It is scarcely necessary to state that
it is now usual to employ shunt-wound machines producing
currents of moderate strength sufficient for a large number of
comparatively small baths coupled in series ; the voltage at the
terminals of the dynamo being chosen according to the number
of the baths to be used, and to the arrangement of the electrodes.
Up to this point all the systems in use are in agreement; but
there are two ways of arranging the electrodes in the bath,
viz. :— The Parallel System and the Series System. In the
former system all similar electrodes in any one bath are, as
the name implies, coupled in parallel, whilst in the latter
system, all the electrodes in the bath are arranged in series.
THE PARALLEL SYSTEM.
The parallel system is the most widely used, and there is
little doubt but that, for practical reasons and in spite of
theoretical objections, it is closely bound up with the future
of copper refining.
Arrangement of the Eleotrodes in the Baths. — In this
system all the electrodes are suspended, anodes and cathodes
alternately, at distances of from 20 to 50 mm. [f in. to 2 ins.]
apart, the first and the last in each tank being cathodes. They
are all suspended, parallel to one another and at right angles
to the longer sides of the tanks. All consist of plates with
arrangements for suspending them and for connecting them up
Digitized by V^OOQ IC
^ ?lfi
ELECTRIC SMELTING AND REFINING.
to the electric mains. The best-known types of electrode and
the process of preparing them must now be described. Figs.
Ill, 112, 113, and 114 illustrate certain forms of anode, and
Fig. 111.
Fig. 114. Fig. 115.
Fig. 112.
Fig. 116.
Fig. 113. Fig. 117.
Typical anodes and cathodes.
Digitized by LjOOQIC
COPPER.
213
Figs. 115 and 116, the forms of cathode corresponding to the
anodes sliown in Fig. 114 and Figs. Ill and 113 respectively.
Casting of Anode-plates. — Instances in which black copper
is bought for the purpose of refining it electrolyticallj, either
directly or after previously refining it by a furnace process, are
becoming more and more rare. The conditions of the electro-
lytic treatment of crude copper are now so well known and are
so well adapted to those of even quite small copper works, that
works which purchase crude copper as such will in no very
distant date be most unusual. Hence, it may be considered
that, under existing circumstances, the anode material is usually
obtained direct from the reverberatory furnace or converter.
And if this is so, and the metal may be run direct into the
form of anode-plates, it is clear that all the extra costs of
re-melting and of treatment in the furnace are obviated.
The anode-plates of Siemens and Halske, and of the Nord-
deuUdie Ajffmeriey which represent the oldest and simplest type,
were so made (Figs. Ill and
114) that they may be hung
directly upon the conductors
carrying the current. In pre-
paring these the copper is
ladled or tapped out of the
reverberatory furnace, and
then, with the help of ladles,
is poured into moulds which,
for the Siemens and Halske
anodes, consist of strong iron
bed-plates with iron frames
projecting upwards from them
shaped according to the form
^nd thickness of the anodes.
One portion of this frame is
attached firmly to the iron
plate (Fig. 117), or is in one piece with it, whilst another part
is removable, and is simply dowelled into the fixed portion
during the period of casting. Moulds such as are used in
running copper bars or ingots suffice for the comparatively
small anodes of the Noi^ddeuUche Affinerie,
Morrow*^ prefers to use lugs of sheet copper cast into the
crude copper plates, instead of the side supports or hooks which
are employed to carry the anodes in the case of the electrodes
above described, and which must be cast of almost the same
thickness as the anodes themselves. The mould used for this
purpose is shown in Figs. 118 and 119.
Anodes of the shape shown in Fig. 112 are chiefly used in
the works designed by Thofehrn, who was at one time a tech-
• Unitefl SUtes Patent 631,471, Aug. 22, 1899.
Figs. 118 and 119.— Mould used by
Morrow for casting anodes.
Digitized by VjOOQIC
214 ELECTRIC SMELTING AND REFINING.
nical assistant with the firm of Siemens k Halske. Thofebm,
believing in the special advantages of using as large baths as
possible, but at the same time not wishing to produce anode-
plates that would be too difficult to handle, employs anodes
the width of which is not appreciably greater than the width of
those used by Siemens k Halske. He then hangs at least
two or three, side by side, by means of hooks, from square bars
placed across the very wide baths. The plates are therefore
provided with lugs or eyes for the hooks instead of with project-
ing arms at the sides. A different type of anode is used in the
works of the Guggenheim Smelting Company (Perth Amboy,
New Jersey, U.S.A.),* where the electrodes are placed more
closely together, with the result that the number of plates, and,
therefore, the output per tank, are increased. These electrodes
(Fig. 113) are provided with the older type of side support only
on one side, whilst on the other side there is an iron pin about
12 mm. [\ in.] in diameter, and about 75 mm. [3 ins.] long,
inserted in the anode to the depth of about 25 mm. [1 in.]. The
cast projecting arm on one side of the anode rests upon the
positive conducting-rod, whilst the free end of the iron pin on
the other side of the plate rests in a small wooden insulating
block which serves as a support for the electrode on that wall of
the vessel. The comparatively small cathode supports rest
(without any iron pin) directly on similar wooden carriers on
the opposite side. In this way it is possible to make the
distance between the electrodes about 20 mm. [J in.], instead of
40 to 50 mm. [1^ to 2 ins.], as is commonly the distance in other
baths. This plant also has a simplified system of connections
which will be described hereafter.
Anode-Casting Machines. — The casting of copper into
anodes with the help of ladles is necessarily comparatively costly
and inconvenient; hence anode-casting machines were soon
devised. Two of these may be described here. That of Hixon
and Dyblie,t shown in Figs. 120 and 121, consists of a number
of frames placed together, and attached to a bottom plate,
exactly like the parts of a filter-press. The bottom plate rests
on a truck, by means of which the whole of the moulds (which
should be capable of containing the entire charge of a converter)
are run up to the converter. On the under side of the bottom
plate are two rails resting on rollers, and between them is a
rack engaging with a pinion actuated by means of a hand- wheel.
With the aid of this device the several moulds may be brought
successively underneath a funnel passing through a cover plate,
so that the copper from the converter may be run into each
in turn.
A very complete casting apparatus, designed by Walker, { is in
* Ulke, Mineral Industries^ 1899, vol. viii., p. 186.
t United States Patent 6.39,270, May 14, 1895.
X Keller, Mineral Industries^ 1898, vol. vii., p. 252. t
Digitized by V^OOQ IC
COPPER.
215
use in several of the large American copper works. The metal
is run from the furnace into a tilting ladle placed before the tap-
hole. The filled ladle is so arranged that the side nearest the
furnace can be raised by means of a lifting device, and the metal
then flows from the spoilt on the opposite side. Thence it passes
through a small runner or deflector, adjustable bv hand, into the
moulds which are carried on arms attached radially to a turn-
table wheel. These arms are so made that they may be adapted
for use with moulds of any required shape, as shown in Figs. 122
and 123. On the side of the turn-table diametrically opposite
Fig. 121.
Hixon and Dyblie's anode-casting machine.
the casting ladle is an elevator by means of which the heavy
metal plates are lifted into the vat-room or on to whatever
apparatus may be provided for conveying them there. By the
time that the plates cast on the furnace side have been brought
round to the elevator by the rotation of the wheel, they are
sufficiently cold to allow of their being [automatically] turned
out of the moulds on to an inclined plane which conveys them
to the elevator. [The bosh in which this inclined plane and the
lower part of the ^levator are placed is filled with water.] They
are also sprinkled with water before they are tipped out of
the moulds.
Digitized by VjOOQIC
216
ELECTRIC SMELTING AND REFINING.
Digitized by LjOOQ IC
COPPEB
Digitized by VjOOQ IC
218
ELECTRIC SMELTING AND REFINING.
Cathode -Plates. — The cathode -plates are always made of
thin sheet copper which, in most refineries, are formed electro-
Ijticallj. For this purpose a thin layer of copper is deposited
on sheet lead or on copper plates, the surface of which has been
prepared by a coating of graphite or by greasing slightly. The
deposited sheet is then stripped from the plate on which it was
formed and cut to the shape and size required for the cathode.
Lastly, it is provided either with holes by which it may be
suspended from the hooks, or with bends to support it on the
conducting bars. At the Buffalo works the matrix plate is
painted with a dilute solution of iodine in naphtha [to prevent
the adhesion of the deposited copper], so that the purity of the
solution is not affected as it is when tallow is employed for
the purpose.
Depositing Vats. — Wooden vats lined with lead are now
Fig. 124. — Section of copper-refining vat, showing anode (Siemens).
almost exclusively used for the electrolytic process. Cement
tanks are too porous for use unless they are varnished. But
with all the acid-resisting varnishes recommended formerly for
use with wooden vats it has been found that, after prolonged
exposure to the action of the constituents of the electrolyte,
Digitized by VjOO^ It^
COPPER.
219
they introduce into the solution soluble substances which act
prejudicially upon the copper that is being deposited.
The other arrangements of the electrolytic baths, and the
subsidiary appliances used, will be best illustrated by a descrip-
tion of actual installations. In the first place, an account may
be given of a plant erected by the firm of Siemens k Halske in
the chemical works of Messrs. Borchers Brothers, at Goslar,
and improved in several important respects by Messrs. K. and H.
Borchers. Thanks to these gentlemen, it is possible here to
publish illustrations of some of the details of the plant.
The anodes were of the original shape commonly used in plant
designed by Siemens <fe Halske, as may be seen at a in Fig. 124.
An account has already been given of the casting of these anodes.
The cathodes always consist of thin copper plate (Fig. 125).
The baths or depositing- cells are wooden vats (pitch-pine), H,
with a leaden lining, which is bent round over the rims of the
Fig. 125. — Cross-section of copper- vat (Siemens) sho^dng cathode.
vats. On this is laid the wooden frame, r, which is saturated
with oil or other suitable substance to prevent it from absorbing
water, and which serves to insulate the two copper bars, + and
- , used as conductors. Before the electrodes are suspended
in place, the lead siphon pipe, x (Figs. 124 and 126), for the
withdrawal of the solution, Ls introduced into the vat; this is
Digitized by
Google
220
ELECTRIC SMELTING AND REFINING.
followed by the wooden table, I, which rests on the bottom
and sapporta a leaden traj, $; the latter is tamed up at its
edges, in order that it may collect the deposit of mnd which
griiuiaallj separates firom the anode.
The anodes hang directly from the sides of the vat, bat mast
be insulated from the - conductors by the india-rubber plates, i
(Fig. 124). The cathodes are suspended by sbeetHX>pper hooks
from strips of wood above. One of the copper strips, bent below
Fig. 126. — Longitudinal section of copper- vat (Siemens).
into a hook, is made longer than the other, so that it may be
folded several times around the wood, as shown in Figs. 124 and
125, and may ultimately make contact with the - conductor,
and be in electrical commtmication with the negative pole of
the dynamo.
In order to maintain the circulation of liquid that is so
essential to the success of the whole process, every bath is
provided with a tube, V (Figs. 124 and 126), to the lower side
of which are attached, at right angles, a number of narrow
tubes, so that the whole arrangement resembles a garden rake.
Each of these rakes is in communication, by means of a rubber
tube, with a main pipe, running along the sides of the vats, and
conveying the electrolyte solution. The supply of liquid to each
Digitized by V^jOO^ It^
Digitized by VjOOQ IC
To face p. ffl.]
Digitized by VjOOQIC
PLATE I.
Ck)pper Refinery— Siemens' System. Scale i : 140.
^sif^*"
Fig. 127.— Sectional Elevation.
■
■
Fig. 129. -Plan.
Digitized by LjOOQ IC
m
Digitized by VjOOQIC
!
COPPER. 221
rake is controlled by a screv clamp, q. The solution is run
away from beneath the slime tray through the siphon, x, into
the channel, Z.
From Figs. 127, 128, and 129, shown on Plate I., will be seen
both the arrangement of the vats and the method of circulating
the solution. In order not to detract from the clearness of these
illustrations the electric connections of the baths are omitted,
but are shown separately in Fig. 130; this figure scarcely
requires further explanation. It will be observed that the
solution is distributed to the baths from the elevated reser-
voir, B, and returns through the channel, Z, to the receivers, G
and D. If all of the liquid is to be discharged, or if one of
the receivers should leak, the solution is drawn off through a
conduit beneath the floor of the room into the tank, £. By
means of the pump, P, the circulation of the liquid can either
be renewed by transferring it from 0, D, or E to the reservoir,
6, or, if too impure for further use, it may be removed to
the regenerating or cementation plant. Regeneration consists
generally in evaporation or crystallisation, if necessary with
I'^^T^: — •---•^---T----::----. j^--^-'-^-r---^---, ^.-■■. .^ —.-.----.p-^^
L.-...j-.jt...l..^.....L..±...l...- .....t....L.../^■..J.....t...j.....-,..i....t....U,^
Fig. 130.— Scheme of electrical connections in copper- vats (Siemens).
the aid of chemical precipitation. Ooncerning the cementation
process, metallurgical handbooks usually give sufficient infor-
mation.
The points to which special attention must be devoted during
the process have been made clear by the work of Kiliani, Wohl-
will, Forster, and others, which has already been described in
full. Reference has also been made to the deterioration of the
deposit caused by the diminution in the proportion of copper
in the bath, which results, under the conditions described,
from an accumulation of non-precipitable metals. So also the
importance of maintaining a brisk circulation of the electrolyte,
of preserving a certain degree of acidity, and of using a small
current-density, were sufficiently emphasised by Kiliani.
Borohers' Improvements in the Siemens-Halske Plant.—
But the means which, until the last few years, had been com-
monly employed to fulfil the necessary conditions, and to avoid
the corresponding sources of failure, were quite insufficient.
The need of completely renewing the solutions when using
the more impure sorts of copper recurred more often than was
to be desired in a continuous process. Many suggestions have,
of course, been made, of which one of the most obvious was the
injection of air into the bath. But the proposal itself was
simpler than the method of carrying it into effect. The direct
introduction of a blast of air into the solution was satisfactory
Digitized by VorOO^ It!
223
ELECTRIC SMELTING AND REFINING.
only for a short time, because it soon led to the production of
a turbid liquid, so that the advantages of this method of stirring
were soon found to be illusory. Yet air was clearly the best
agent for the stirring, and, up to a certain point, for the purifica-
tion, of the liquors, if it only could be satisfactorily applied, and
the idea of using it must undoubtedly be considered one of the
most important discoveries in connection with the refining of
copper. The merit of solving the problem belongs to Messrs.
K. and H. Borchers, of the above-mentioned firm of Borchers
Bros. ; and this invention, it should be observed, is not patented.
The method of carrying it into effect will be first described. A
]j4
\
^W^
iC
Fig. 131.— SiemensBorchera* copper-vats (longitudinal section),
bath fitted with Borchers' apparatus is shown in Figs. 131 and
132. On comparing these with the previous figures it will be
seen that there has been practically no change in the arrange-
ment of the electrodes. But there has been added a wide leaden
pipe, 6, which passes downwards from the upper surface of the
liquid to the very centre of the space beneath the mud-depositing
tray, 8 ; and within this pipe is a glass tube, g, drawn out to a
fine jet at its lower extremity. The glass tube is passed through
a stopper in the leaden cap, dy by which the tube, 6, is sur-
mounted, and is so arranged that it may easily be raised or
lowered in position. A current of air is continuously forced
Digitized by V^jOO^ It^
COPPER.
223
through the glass tube into the column of liquid in the pipe, b,
with the result that the air, dividing itself into small bubbles
as it rises to the surface, lowers the specific gravity of the
contents of the tube, b, and causes a portion of the liquid to
overflow into the vat, where it mixes with the bulk of the
electrolyte. Meanwhile, solution from the lowest level of the
bath is constantly entering at the bottom of the tube, b. The
Fig. 132. — Siemens-Borchers* copper-vats (cross-section).
aeration of the electrolyte is thus effected very quietly and
uniformly, whilst at the same time the method of applying the
air affords the most practical and economical means of circulat-
ing the liquid. The process would therefore commend itself to
the notice of the refiner on the latter ground alone, even if we
disregard the advantages to be derived from the chemical action
of the air, which, however, must not be under-rated.*
♦[According to the Engineering and Mining Journal, 1896, vol. Ixii.,
p. 603, it is found desirable, in practice, to assist the circulation of the
electrolyte for two to three hours in the course of a day, by causing a
general circulation from vat to vat, in addition to the air-mixing accom-
plished by Borchers' device. — Trahslator.]
Digitized by V^jOOQIC
224 ELECTRIC SMELTING AND BEFININQ.
A glance at Figs. 131 and 132, suffices to show that a very
inconvenient portion of the older apparatus — the ''rake" for
distributing the solution — ^has been dispensed with, and that
the side channels for the conyeyance of the liquid have also
disappeared. These two attachments of the original installa-
tion had rendered cleanliness in working the process impossible,
owing to the constant and unavoidable splashing of the solution;
while at the same time thej prevented ready access to the
electrodes, and were, in fact, the chief sources of trouble in the
whole installation on account of the constant supervision that
they demanded.
Details of the Siemens-Borohers' Installation.— In the
plant, as at present used, both the filling and discharging of the
baths are effected by a single pipe, R (Fig. 131, and Plate II.,
Figs. 133, 134, 135), the former being done once for all at the
beginning of a refining operation, and the latter being necessary
only when the impure liquors have to be run off at the end, or
as required. This is in striking contrast to the older system,
which necessitated a constant circulation of the electrolyte
through the whole series of vats. Every individual vat is
connected with the main pipe, R, by means of a siphon, N,
attached to it by a rubber tube, S (Fig. 132). During the working
of the process this tube, S, is closed by a screw clamp, so that
there may be no electrical communication between the baths.
Plate II., like the previous plate, is concerned mainly with
the method of circulating the solution, and with the general
arrangement of the electrolytic plant, the details of bath con-
struction having been shown on a larger scale in Figs. 131 and
132. In the new installation a compressed-air chamber, D,
effects the conveyance of the bath-liquor. In starting work the
electrolyte, placed originally in A, is allowed to run into the
pressure vessel, D, whence it is forced into the distributing vat,
Y, so that it may flow to the vats through the pipes, R, and the
siphons, N. The liquid may be discharged from any given vat
during the electrolysis by drawing it into the pressure kier, D,
through the siphons, N, and the mains, R, so that it may be
treated either for the recovery of its salts or by cementation or
regeneration. Any solution which may escape in consequence
of leakages in the vessels or conveying pipes, or through irregu-
larities in the working of the vats, is conducted through the
drains, G (placed beneath the flooring), into the tank, B, whence
it may be conducted to the pressure chamber for re-distribution.
The vats, C, for the washing of the anode slimes have also found
a place in this room under the present arrangement.
Siemens & Halske recommend that the solutions shouM be
warmed, the conditions being then more favourable for the
precipitation of antimony and bismuth compounds.
Forifloation of Old Iiiquors fcom Bismuth and Anti..
t
Digitized by V^jOOQI?
\To ^mee p. ttU,
Digitized by LjOOQIC
Digitized by VjOOQ IC
COPPER.
225
mony. — If, in coarse of long use, the electrolyte should become
saturated with compounds of these two metals, it must be run
out of the vats into a specially prepared tank, and be there
warmed, treated with air forced through it by means of a steam
injector, neutralised with copper oxide or other suitable basic
copper compound that may be at hand and filtered. After
reserving a portion of the solution so obtained for working up
into copper sulphate, the remainder is returned for use in the
refinery, afber adding sufficient sulphuric acid, and bringing it
to the normal strength, for the work.
Beduction in the Cost of Eleotroly tio Befining.— Finally,
a by no means unimportant gain in using the new process, is to
be found in the fact that the current-density may be increased
from 30 to 100 amperes per sq. metre [9 3 amperes per sq. ft.]
without detriment to the texture or surface of the deposited
copper. The full meaning of this advantage will be appreciated
after the perusal of the following translation of a private com-
munication from the firm of Siemens <k Halske to the author : —
" The possibility of considerably increasing the current-density
has the great advantage that, for the same outturn of refined
copper, the floor space occupied by the baths, and the quantity
of copper, silver, and solution locked up in the vats, will be re-
duced to about one-third, and the item of wages will be con-
siderably lower. On the other hand, the power required will be
greater; in spite of that, however, there will be a considerable
nett saving in the working cost. In the following table the
estimated cost of working at 30 amperes and 100 amperes per
sq. metre are placed side by side for comparison, the calculation
being based upon the ordinary conditions of work in Germany.
Where other conditions prevail, the necessary alteration in the
figures may readily be made; but in any case the difference
between the costs of working by the old and new processes will
remain very much the same as is here indicated : —
COMPARATIVE WORKING COST WITH HIGH AND LOW
CURRENT-DENSITY.
Daily working oost for an output of 1 ton
of reiined copper per diem.
Cost of power (I.H.P. hour = 5 pf.
[id]). ......
Wages,
Interest on copper looked up (5 /o),
AmortiBation of the electric plant (10%),
Cost of heating the baths, 250 kgs.
[5 cwts.] ooal, ....
Cost of regenerating the electrolyte,
Or, taking 1 mark = 1 shilling, .
Former current- Present current-
density 80 amps, density 100 amps
per sq. m. I per sq. m.
M.17
15-60
8-30
74-90
£3, 148. Ud.
M.30
15
4-80
416
68 95
, £2, 188. lid.
15 T
Digitized by VIjOOQIC
226 ELECTRIC SMELTING AND KBFININO.
*' From this table it is evident that the actual cost of refining
has been reduced by about 20 per cent, through the introduction
of tlie new process ; and since the interest, both on the capital
sunk in buildings and site, and on the silver in the anode
copper, will also be reduced, the actual comparison is even more
in favour of the application of high current-density than would
appear from the above table.
"In order to introduce the improved process into existing
works, either the present baths, after the necessary alterations
have been made in them, may be used to produce a higher
output of copper by increasing the power of the engine and
dynamos ; or the same power may be employed with a smaller
electrolysing plant and a somewhat diminished outturn, but
with a higher efficiency ; in the latter, about three-quarters oi
the area originally occupied by the baths will be set free, and
80 rendered available for other purposes. Generally, however,
the former alternative will prove more advantageous."
The firm of Siemeus & Halske, in a communication to the
Zeitachrifi fUr Elektrochemie, have claimed for Werner von
Siemens the invention of the application of air to the circulation
of solution, above ascribed by Borchers to the members of the
firm of Borchers Bros. They state that " this kind of air circu-
lation was discovered by Dr. Werner von Siemens in 1884 or
1885, and was then used for the raising of water. The first
plant installed was shown in actual use to a number of
engineers, and was known as a 'Geyser Pump.' A double
tube was sunk into a well-shaft, and air was blown into the
inner tube by means of a steam engine working as an air-
injector pump. Siemens & Halske used this means for cir-
culating the solution at the beginning of 1886 in an electrolytic
copper refinery which they were laying down in the copper
works of 0. Heckmann. The same firm started the copper
refinery in Borchers Bros.' works in the early part of 1891 ; but
the circulation was in this case effected by the overfiow of the
solution from tank to tank and not by means of air. Borchers
Bros, themselves introduced the system of air circulation into
their plant afterwards; but, as applied by them, it did not
appreciably difier from that in Heckmann's works. At a later
date the arrangement was considerably improved by W. von
Siemens, and wa<«, after his death, applied to electrolytic baths
of very difierent descriptions."
Borchers writes that while obviously there cannot be the
slightest doubt on the accuracy of the above statement, it must
not be overlooked that the plant actually installed by Siemens
^ Halske in the works of Borchers Bros, was not provided
with air circulation, but that, on the contrary, the latter firm
themselves recognised the advantages of, and applied, the
system, and that they did so without any suggestion from the
original designers of the plant
Digitized by V^OO^ It!
COPPER. 227
The method of work and the construction of the apparatus
above described undoubtedly possess all the advantages which
may reasonably be expected from a well-planned electrolytic
installation — namely, simplicity of arrangement, convenience of
access, and control in all parts, cleanliness in operation, and
exclusion as far as possible of all intermediate work of a dis-
turbing character. It might have sufficed to limit the account
of copper refining to that already given, and to recommend the
system of Siemens-Borchers for universal application, but certain
other works which differ in many ways from those described
above must also be descnbed.
[Before proceeding to the consideration of these processes,
however, it may be worth noting that in American practice
a much higher current-density is used than that employed in
Europe ; it is rarely much lower than 10 amperes per sq. ft.,
and, in some cases, is nearly double that amount. In conse-
quence of this the turnover is greater, and the cost of refining
is correspondingly reduced, inasmuch as the fixed charges are
distributed over a much higher tonnage of refined material, and
the interest on capital locked up in the copper in the vats is
also diminished. The whole process is worked at a lower cost,
and this helps to compensate for the higher charge for wages.
Thus, although the average wages at the Anaconda Works (see
p. 232) is about $3 per diem, the cost of refining (including tnat
of treating the silver mud) is only $14 (£2, 18s.) per ton, whilst
at Perth Amboy (New Jersey) it is stated (exclusive of office
expenses) to be $9 (£1, ITs.) per ton, as against 59 marks
(£2, 19s.) per ton in the table given on p. 225. The charge for
labour at Perth Amboy is less than half that at the Anaconda
Works, but it is still nearly double that in Germany. Hence
the gain in applying the highest current-density that can be
safely used, without endangering the quality of the deposited
copper, is emphasised as strongly as by the figures quoted in the
above table {loc ek). — Translator.]
I^Sohneider and Saontagh System. — A system of air circu*
lation similar to that of Borchers, has been patented in the
United States by Schneider and Szontagh.'*^ The method of
introducing the air into a vertical pipe communicating with the
electrolyte tank above and below is practically the same ; but,
instead of taking otf the liquid from beneath the centre of a
tray, it draws it through perforations in a horizontal pipe
running along the whole length of the vat at one of the bottom
angles, whilst the solution overflowing from the upright air-pipe
passes through a second horizontal pipe, which is placed along
the top of the vat on the opposite side. From this pipe it is
delivered into the tank through perforations pierced between
each pair of electrodes, and increasing in diameter as they are
♦[U.S. A Patent 663,093, June 30, 1896.]
Digitized by LjOOQ IC
228
ELECTRIC SMELTING AND REFINING.
farther removed from the vertical pipe, in order to secure a
greater uniformity of flow. The system is in use in the large
refinery of Messrs. Guggenheim at Perth Amboy, New Jersey.
— Translator.]
The Thofehm Process. — According to the account of the
Thofehrn process published by Hering,* it may be distinguished
from all others by the use of very large baths, for which either
wood lined with lead or concrete painted with tar may be em-
ployed. The baths have a width of about 2 metres, a length of
3 metres, and a depth of about 1*5 metres [6| ft. x 10 fb. x 5 ft].
The form of the anodes is the same as that adopted in the
Moebius apparatus for treating silver,! and the method of
suspending them in the bath is also the same — viz., by hanging
several small plates from a metal bar. In Thofehm's baths each
anode rod carries three plates 0-6 metre long by 0*6 metre wide
by 002 metre thick [2 ft. x 2 ft. x f in.]. The cathode coDsists
of copper plates, of which six are hung on every rail, each one
being 06 metre long by 0*2 metre wide by 00002 metre thick
[2 ft. X 8 ins. x 0-008 in.]. A diagrammatic sketch of the posi-
tion of the conductors used in conveying the current to the
supporting rods for the electrodes is given in Fig. 136. It will
t
T
Fig. — 136. — The electrical connections of Thofehm's apparatus.
be recognised at once as an old arrangement which had been
applied in plant laid down by Siemens <b Halske before any
account of Thofehrn's process was published. As an electrolyte
for black copper anodes with a current-density of 30 amperes
per sq. metre [2'8 amperes per sq. ft.] he recommends an
aqueous solution of 15 per cent, of copper sulphate and 5 per
cent, of sulphuric acid ; or with 50 amperes per sq. metre [4*6
amperes per sq. ft.], one of 20 per cent, of copper sulphate and
5 '5 per cent, of sulphuric acid; whilst for bessemerised copper
and a current of 60 amperes per sq. metre [5*6 amperes per
sq. ft.] he prefers a bath containing 25 per cent, of the copper
salt and 6 per cent, of the acid. The circulatiom of the solution
from bath to bath is effected with the aid of siphons, which,
however, afford no security for the uniform circulation of the
electrolyte through the large baths themselves. An interesting
account was given by the Engineering and Mining Journal, in
•Herina, Berg- und HiUUnmanni9che Zeitung, 1893, voL lii., p. 53;
Revue Industridle, 1892, p. 24.
tSee Chapter on Silver,
Digitized by VjOOQ IC
OOPPEB. 229
1896, of a plant laid down by Thofehrn in the works of the
Anaconda Mining Oompany, and designed to refine 150 tons of
copper per diem.
The engine-house contains an ordinary double-cylinder engine
of 800 H.P., a Westioghouse compound engine of 400 H.P., and
two triplo-expansion engines, each of 900 H.P. Each of the
two latter engines is direct-coupled to two dynamos. There are
thus 3,000 H.P. provided, but this includes the reserve. The
first-named engine is of the Oorliss type, and drives the dynamos
of the old plant by belt transmission. For the work of electro-
lysis there are provided — three Westinghouse dynamos, all
belt<lriven, two of 220-kilowatt capacity each, and one of
270 kilowatts; and four Westinghouse dynamos, each of
270-kilowatt capacity, two coupled to each of the triple-expan-
sion engines. These dynamos afford, therefore, 1,790 kilowatts,
which corresponds to 2,400 E.H.P. One of these dynamos
{220 kilowatts = 295 E.H.P.) with its corresponding steam
engine forms the reserve; and can be connected to any one
of the six circuits in the event of one of the other dynamos
requiring repair. There are, in addition, two dynamos for
lighting and power-transmission, one of them being held in
reserve. An air pump absorbing 30 H.P. is used to drive the
acid pumps. As the whole plant is constructed to deposit
170 tons of copper a day, it appears that the expenditure of
work is at the rate of 17*5 H.P.-hours per ton of copper,
inclusive of all mechanical work which, as far as possible, is
done by electrically-driven machines.
The refinery itself is divided into two blocks, separated by a
free space of about 100 ft., in order to minimise the fire risks.
The buildings are constructed of wood covered with corrugated
iron ; each building has a floor space of about 6,500 sq. yards
and contains about 600 tanks, each measuring 2'5 metres in
length by 1*5 metres in breadth and 1 metre in depth [8 ft.
3 ins. X 4 ft. 7 ins. x 3 ft. 3 ins.]. In the construction of these
vats, a wooden frame, sufficiently large for 10 vats, is first put
together, this is lined witli planks, and so divided up as to form
10 compartments separated from one another by sufficient air
space. Each of these rows of 10 vats is mounted on its own
foundations, separated from the next rows and from the working
floor. All joints in the wood are well caulked with insulating
material. The vats are lined with lead and carry the conduc-
tors, insulated from the lead lining, on the top of the side walls.
The baths are connected according to the system known as the
Siemens system. The electrodes are suspended by copper hooks
from flat transverse iron bars which rest on the longitudinal
conductors. The actual number, size, and disposition of the
electrodes are not stated, but from the illustrations given (cf.
Pig. 136a) it would appear that to every vat there are 16
Digitized by V^OO^ It!
230 ELECTRIC BH^LTINO AND REFINING.
cathode rods, and 15 anode rods, each supporting two or more-
strip-electrodes. Two compartments, each containing 10 rows
of 10 tanks, or 200 tanks in all are coupled in one circuit; and
each of these circuits is arranged in one hall, and thus, with the
necessary mechanical appliances and general equipment, forms a.
system that is quite independent of all the other systems. All
the baths in any one system are coupled in series, and it has
already been stated, in the preceding paragraph, that there are
six of these systems. A reservoir is provided for each row of
tanks from which the circulation of the liquid is ensured, the
solution flowing through the 10 vats in succession and thence to
a collecting tank, whence it is again pumped with the help of
pressure kiers to the distributing reservoir.
In the three systems first installed, the introduction and
removal of the electrodes was effected by means of overhead
pulleys. The full load of a tank may amount to about 4 ton&
of copper, the whole of which is introduced at one time. In
the three systems more recently laid down, electric cranes-
[overhead travellers] are provided for this work. The raw-
and fine-copper plates are conveyed in the vat-room and through
the works by means of an electric tramway of 500 mm. [20 ins.]
gauge. The rails are brought in to the working floor of each
hall, and other rails are run beneath the depositing tanks for
the purpose of conveying the anode-mud to the silver refinery
attached to the works.
Method of Work. — The anode-plates are weighed when first
brought into the refinery, and are then suspended by copper
hooks from the flat transverse iron bars already referred to,
which in turn are lifted by means of the pulleys or traveller
and dropped into position on the tanks. The cathodes are
similarly prepared for suspension between the anodes. Finally
the solution is run into the tank, which is then connected up in
the electric circuit. The work of emptying, cleaning up, and
re-charging a vat requires a little less than an hour. It is
scarcely necessary to point out that the process of emptying a
tank is similar to that of charging it, but in inverse order.
The anode material is black copper containing 98 per cent,
of copper, with small quantities of arsenic, antimony, selenium,
tellurium, iron, and lead, and about 0*35 per cent, of silver, and
O'OOl per cent, of gold [110 ozs. Ag and ^ oz. Au per ton].
In order to control the work of depositing, leads are taken ofiT
from every set of five tanks to a measuring apparatus which
automatically registers the reading once an hour. To effect thia
each series of five tanks is connected up to a kind of dynamo
commutator, in such a way that the opposite poles of the same
series of ^ve vats are joined up to radial copper plates at the
opposite ends of one diameter of the commutator. These plates
are all insulated from one another. A yoke-piece with two
Digitized by V^OO^ It!
232 ELECTRIC SMELTING AND REFINING.
brushes, diametrically opposite, is revolved around the com-
mutator once an hour; and the brushes are connected to an
automatically-registering voltmeter, so that each series of five
vats is individually connected up to the latter instrument, and
the slope of potential between the terminals of each of these
series is therefore registered serially once every hour. The
manager is thus able to detect and to locate precisely any
irregularities that may be taking place.
Conditions of Work, — The refinery of the Anaconda Copper
Mining Company yields from 100 to 120 tons of copper per diem.
The remainder of the copper — amounting to about 80 to 100
tons — produced by this Company is refined in Baltimore, so that
the daily outturn averages about 200 tons in all. The refinery
above described is now so arranged that, in case of need, the
whole of the daily production of the Anaconda Company
could be treated there, with the addition only of the necessary
dynamos.
In regard to the cost of the work of refining in Anaconda, it
may be reckoned that the wages average about 12s. ($3) for each
man daily. The fuel used had formerly cost from 22s. to 24s.
($5.50) per ton; but the new boilers have been arranged to run
with a coal costing 8s. to 8s. 6d. ($2) per ton. Sulphuric acid
costs about £10 per ton. It may be generally assumed that, in
Anaconda, all expenses are about twice as high as in the in-
dustrial centres of the Eastern States of America. The total
cost of refining is placed at $14 (nearly £3) per ton of refined
metal.
There are 120 men employed in the works, including the
foremen, assayers, and clerks. The foregoing illustration (Fig.
136a) serves to show the economy of labour in a modern plant.
The older existing buildings were not strong enough to carry all
the modem labour-saving appliances, but the new plant was put
down according to plans prepared by Thofehrn. Both plants
turn out the same quantity of copper — namely, about 50 tons
per diem — but in the old plant 50 men are employed as against
25 in the new portion, and this must certainly efibct a great
economy in wages.
The purification of the electrolyte is only effected with a view
to keeping the proportion of the anode impurities passing into
solution within certain limits ; but the writer of the account of
the works gives no indication as to the method employed beyond
saying that it is simple, and calls for the use only of inexpensive
chemicals in addition to air.
The monthly production of gold is 46 kilogrammes, 0-950 fine;
and of silver, 10,886 kilogrammes, 0-999 fine.
The washed anode-mud is boiled with acid by means of steam
after the pieces of copper have been removed from it, and is then
tiltered and washed. Compounds of arsenic and antimony are
Digitized by VorOO^ It!
COPPER. 233
next brought into solution in a second series of vessels, but the
method employed for this is not described. The mud is now
filtered again and washed, and is then dried in large cast-iron
pans, melted in a reverberatory furnace, and cast into small bars,
which are finally treated by the sulphuric acid process.
The current-density used is approximately from 100 to 200
amperes per sq. in. [10 to 20 amps, per sq. ft.] of cathode surface.
The copper produced is melted and run into bars of suitable
shape for the wire-bar rolling mill.
[This is effected at the cost of about $4 per ton; and this
metal, in the form of hard drawn wire, containing, as it must, at
least a trace of cuprous oxide, is found to have an electrical
conductivity of 98 (Matthiessen's standard), and a tensile
strength of 64,000 to 65,000 lbs. per sq. in.; the number of
twists obtainable in 6 ins. of No. 12 wire is 80; elongation,
1 J per cent.]
The design of the plant as a whole is, doubtless, good, although
the advantages of using such wide baths may not be apparent,
and the design itself may not give much evidence of great
originality. The arrangements of other electro-metallurgists,
especially of Siemens and Halske, and Moebius, are observable in
many places.
Thofehm's New Frooess. — The same number of the
Engineering and Mining Journal {loc. cit.) describes a new
process in which the copper is deposited upon a long hollow
cylindrical cathode 8 ft. long and 3 ft. in diameter, which is
immersed in the electrolyte, and is revolved at a low speed,
whilst the copper is deposited upon it with a current-density of
50 to 100 amperes per sq. ft. During the time of depositing,
numerous jets of the electrolyte solution are caused to play
under pressure upon the sur&ce of the cylinder.* It is stated
that the crystals of copper are deposited in the form of micro-
scopic octagonal hair-like filaments, which become felted and
compressed by the jets of liquid ; and that the cylinders of
* [A similar device has been adopted by Graham (English patent 986,
Jan. 14, 1896). According to his specification, a nearly saturated solution
of copper sulphate, containing 5 ozs. of strong sulphuric acid per cb. ft. of
water, is used as electrolyte, and is stored in reservoirs placed at a height
of^ from 1 to 2 ft. above the electrolytic tanks, thence it is conveyed to
§<in. jets that deliver it in a stream u(K)n the surfcMse of the cathodes, which
are placed at a distance of 1^ ins. from the orifices of the jets. It is stated
that a current of 300 amperes per sq. ft. may be employed with safety at
all points within the sphere of influence of the jets ; beyond this area, it is
obvious that the deposit would be pulverulent and useless. The radius of
the protected area is found to be 5 ms. , and the number and disposition of
jets employed must be arranged accordingly. The anode should have a
surface area eight or nine times greater than that of the cathode, and its
shape should preferably be that of a corrugated spiral or of a grid, through
which the electrolyte streams are directed upon the cathodes. The edges
of the latter are protected by an insulating material. — Translator.]
Digitized by V^OOQ IC
234 ELECTRIC SMKLTINO AND REFINIKG.
copper when they have been deposited an inch thick and have
been detached from the cathode cylinders may be rolled direct.
It is said that the wires are over 15 \ier cent, stronger than
those prepared by the old electrolytic refining process followed
by fusion ; and that the expense of such an electric refining of
the metal to produce the bars required for the rolling or wire-
drawing mill is )$16 (£3, 6s.) per ton.
One distinct advantage in the use of these jets is obviously to
be found in the certainty that the electrolyte at the surface of
the cathode is constantly and thoroughly renewed, so that a very
much greater current-density may be employed than would
otherwise be possible. No details whatever are given concern-
ing the process, and it is therefore impossible to criticise it. It
may be noted that the same paper states that white metal anodes
with 75 to 80 per cent, of copper have been employed in thia
process with success, but no hint is given as to the manner in
which difficulties that have baffled the attempts of previous
inventors have been obviated. — Translator.]
The Baritan Copper Works Plant. — Finally, reference may
be made to an account given by Addichs**^ of the electrolytic
installation of newly-erected works belonging to the Baritan
Copper Works, on the Raritan river, at Perth Am boy, New
Jersey [U.S.A.]. The works are designed to produce from 5,000
to 6,000 tons of electrolytic copper per month, in addition to
from 7,000 to 8,000 tons refined by the furnace process. The
electrolytic plant is contained in a building measuring 200 ft. by
600 ft., and includes 1,600 depositing vats in which the electrodes
are arranged on the parallel system, whilst the baths are placed
in series. 180 of tliese vats are used only for the electrolytic
production of the original cathode-plates. These are prepared by
depositing copper upon rolled copper plates about 4 mm. [0*15 in.}
thick. In order to facilitate the removal of the deposit the
plates are smeared lightly with tallow, and are then clamped in
wooden frames, so that the edges may remain free from deposit.
In about 36 hours the deposit will be sufficiently thick to allow
of it being stripped from both sides of the plate ; the sheets aro
then sheared otf straight at the edges and bored by suitable tools
to take the suspension hooks. The cathode-))]ates so deposited
are then carefully straightened by striking with wooden mallets
or blocks, and are suspended in the electrolysis cells by means of
copper hooks from copper rods, which serve to make contact
with the conducting bars conveying the current. They remain
ill the bath for seven days, by which time they are sufficiently
thick for further treatment. The anodes are made according to
the well-known pattern, with side supports.
Each vat contains 22 anodes and 23 cathodes, and the current-
density used is about 150 amperes per sq. metre [15 amp. per
* Mineral IndxiAtry, 1900, vol. ix., p. 261.
Digitized by LjOOQ IC
COPPER. 235
sq. ft.]. The main conductors can carry a current of 4,000
amperes, and consist of copper bars 1^x4 ins. in cross-section.
The depositing vats are arranged 400 in series, each series divided
into two for the purpose of circulating the electrolyte, which
therefore passes through 200 tanks in succession. In order to ,
facilitate the circulation of the liquid the tanks are placed in
double-terrace fashion, sloping, /^-like, away on each side from a
central feed-pipe.
In order to regulate the proportion of copper in the electrolytes
each group of 400 tanks has 8 small tanks connected in series
with it. All these smaller vats are placed in a small annexe at
one side of the main building, and are used for electrolysis with
lead anodes. It is well known that the electrolyte usually
becomes richer in copper as the work of deposition [with copper
anodes] proceeds, and these small tanks with lead anodes serve
to reduce the excess of copper. But as in doing so there must
be a brisk evolution of gas, the tanks used for tlie purpose must
be placed in a separate and well-ventilated room.
In the main depositing plant the electrolyte is maintained at
a temperature of 49° C, and this is accomplished by pre-heating
the liquid with steam in the distributing reservoir by means of
a coil of lead pipe. Special vessels are provided for the separation
of the tallow from that part of the solution which is used for the
electrolytic production of the cathode-plates. Each of these
vessels contains a partition which does not reach to the bottom.
The solution flows through the space beneath this partition,
leaving the tallow in the first compartment. The solutions
remain in circulation for several months before a process of
purification becomes necessary. The method' of working up the
impure electrolytes will be referred to hereafter.
Conneotions of Eleotrodes. — In treating of the various
forms of anode reference was made to a system of connecting up
the baths, the advantages of which must be at once apparent.
Every two baths are placed side by side, so closely that the
cathodes of the one bath and the anodes of the other are suspended
from a common conducting bar. [Compare Fig. 149.] Apart
from the fact that in this way one conductor is dispensed with
for every two baths, the middle bar may be of considerably
smaller section than the leads conveying the main current to and
from the baths, for it conveys the current direct from the
cathode of the first bath to the anode of the next. In the works
of the Guggenheim Smelting and Refining Company, where the
side conductors are \\ ins. thick, the middle strip is only \ in.
thick, and is slightly narrower than the others. This system of
connection was adopted in connection with the so-called Marchese
system at Casarza.^
* Compare p. 247.
Digitized by VjOOQ IC
236
ELECTRIC SMELTING AND REFINING.
SERIES SYSTEM OF REFINING.
The Stalmann Process. — The onginal voltaic battery has
been made the foundation of another system of arranging
electrodes, which has come into use in various modifications,
and in some cases with the stated object of economising cathode
plates. Schnabel'"' describes one of these methods (in which,
however, special separate cathode-plates are used), and states
that he has seen it in use in the Anaconda Works at Montana,
U.S.A. Without this testimony it would have been very
difficult to believe that this most inconvenient arrangement
should have found any favour in practical work ; but as it has
Fig. 137. Fig. 138.
Stalmann's electrode connections.
Fig. 139.
been shown to be possible, it will be necessary to describe the
apparatus and the method shortly at this point, the account
being taken from the patent specifications of Stalmann, t whose
process it is.
Stalmann couples the electrodes of each individual bath in
series, suspending an anode first, and then cathodes and anodes
alternately, united in pairs, until at last a cathode-plate ends
the whole series. The first anode and the last cathode are
connected immediately with the main leads of the circuit, or
are joined up in series with other baths similarly arranged.
The details of the electrode connections in the bath are shown
in Figs. 137 to 139. In arranging the first pair of electrodes,
♦ C Schnabel, Handbueh der MetdUhiiU^ikunde, vol. i., p. 270 (18d4).
t U S. A. Patents 467,360 and 467,484, .Jan. 19, 1892.
Digitized by LjOOQIC
COPPER.
237
either the cathode-plate, k^ is fieistened directly to the anode, a
(Fig. 137), or each anode-plate, a, is connected with a cathode-
plate, k^ by means of a wire or by the short copper bars, v, lying
on the rim of the vat (Fig. 138), or, finally, the anode, a, is
united to the cathode, A;, in the manner shown in Fig. 139, with
a plate, t, of insalating material separating them, so that only one
side of each plate comes in contact with the solution. The last-
Fig. 140. — Stalmann's copper vat (cross-section).
named method is the latest. The pairs of electrodes are suspended
in the baths after the fashion adopted in the older systems. If
the electrodes are not attached by screws to supporting arms
as shown in Fig. 138, the anode-plates, a, are cast with pro-
jecting lugs, a; (Fig. 140), by which they are supported on the rim
of the bath. The electrolyte vats (H and H^) are double, one
being placed within the other, whilst the intervening space is
filled up with paraffin wax, tar, or
other similar material ; they are
ffflIB
n.
I I I
I— Ir-J
Fig. 141.— Scheme of electrical
connections in Stalmann's
vat.
made of wood, and the inner vessel
is provided internally with a series
of wooden studs, F, on its side
walls to prevent any displacement
of the electrodes. The connections
of the electrodes and conductors
will be understood on reference to Fig. 141, in which a is
the first anode, k the last cathode, and aJc the intermediate
double electrodes. It is not easy to understand from this
patent specification what advantages Stalmann's arrangement
possesses oyer the other systems in use.
Digitized by VjOOQIC
i^d8 ELBCTRIC SMELTING AND REFINING.
The Haydexiy Smithy and Bandolf Processes. — Hajden*
simplifies this apparatus by omitting all the cathode -plates
except the one connected to the negatiye wire of the generator.
He joins np the first anode-plate to the positive conductor from
the djnamo, and suspends behind it a series of crude copper
plates insulated from it, and from one another, and finally con-
nects a plate of pure copper to the negative lead. The pure
copper is therefore deposited on those sides of the intermediate
plates that are turned towards the first anode, whilst the metal
is at the same time dissolved from the other sides, which face
towards the last negative plate. But, unfortunately, the crude
copper plates, which are always of cast copper, are never quite
uniform in structure ; or, even if the electrodes could be obtained
perfectly homogeneous, certain irregularities due to the clinging
of the insoluble residue to the surface and the like, would be
unavoidable ; there must, therefore, be a more rapid solution of
the metal at some places than at others. Oavities are thus
formed, which, in course of time, extend to the pure metal
f-..^...-...-,-^-j— , deposited (by this time to some
T T I T I thickness) on the other surfEU^e;
{"T^^ that the pure copper must then be
^ ^ -.1.._.|_..J....Lj dissolved is a fact that does not
Fig. 142.-Scheme of electrical need further emphasis. Fig. 142
connectioiiB in Hayden's vat. shows the electrical connections in
the Hayden bath, in which a is
the anode- and k the cathode-plate. The intermediate plates,
z, serve as cathodes on their left-hand and anodes on their
right-hand surfaces.
[The Hayden process, which was introduced in 1886, may perhaps
be said to have been anticipated in principle by Hugon in 1884
and by Farmerf in 1885. The Hayden system was at one time
very largely employed in America, and although in most cases
it has been relinquished, it is still used in the large works of the
Baltimore Electrical Refining Oo. It is said, however, that
even here a large quantity of scrap copper is produced in
working it ; and this is of necessity a serious disadvantage.
Difficulty is often found in series processes in stripping the
deposited copper from the residue of the previous anode
material ; the labour involved in this process adds considerably
to the cost of refining, and occasionally it has even been found
more economical to remelt the whole cathode with the next
charge of anode metal. To obviate this, Stalmann has proposed
sticking a sheet of paper to the back of each anode, fastening
a few rivets through, and blackleading the surface of the paper.
By this arrangement the copper would be deposited upon
the paper instead of upon the surface of the anode copper.
— Translator.]
• £higineering and Mining Journal (New York), 1892, vol. liv., p. 126.
t [U.S. A. Patent 322,170, July 14, 1886.]
Digitized by V^jOO^ It!
COPPER. 239
If Fig. 142 be tamed through a right angle, so that a is
placed at the top and k at the bottom of the illustration, and if
•canvas diaphragms be imagined stretched between the electrodes,
« fair mental pictnre of the Smith system* will be obtained.
The object of the canvas diaphragms is to retain the deposit of
anode-mud, which would otherwise fall u])on the surface of the
cathode below and render the deposited metal impure.
If now the same figure be turned in the opposite direction, so
that ^ is at the top and a underneath, the principle of the
Bandolf arrangement f is sufficiently clear. The copper ions
in this case travel upwards, and as the impurities, therefore,
remain beneath, there is no need for the diaphragm which was
employed by Smith. The circulation of the solution is effected
in a horizontal direction.
In both these latter forms of construction the discovery of the
cause of any accidental disturbance of the operation is rendered
much more difficult. With the vertical suspension of the elec-
trodes, however, such a disturbance is readily found, and very
often is easily remedied. Neither of these methods, therefore,
appears to afford any grounds on which its adoption could be
recommended.
[The Multiple and the Series Systems of Beflning. — The
usual processes of copper refining, such as those of Siemens-
Halske, Thofehm, and others, in which all the anodes in any
one tank are hung in parallel circuits, are now classed as
belonging to the Multiple system, to distinguish them from the
Hayden and similar processes, in which the anode and cathode
pairs within the vat are connected in electrical series. The
latter arrangement is accordingly known as the Series system.
It must be remembered that these terms refer only to the
disposition of the electrodes in a single tank, and do not in
any way relate to the connections of the different tanks in the
installation, which must be so grouped in a combination of series
and parallel, as to give the maximum economy with the current
conditions available, taking into account the interest upon the
capital locked up in the copper contained in the baths, as well
as in the plant itself. The selection of the above-mentioned
names is, therefore, somewhat unfortunate, as it tends to con-
fusion. The amount of copper deposited under the Multiple
system should not be less than 95 per cent, of that which is
theoretically podsible, whilst in the Series processes a higher
RM.F. is required for each tank, and there is a greater loss by
short circuiting through the mud upon the bottom of the bath
and through the walls of the tank itself, so that the electrical
efficiency is reduced to from 85 to 90 per cent. The capital
represented by the copper electrodes is also greater in the case
* Engineering and Atini7ig Journal^ 1892, vol. liv., p. 126.
t U.S.A. Patent 614,275, Feb. 6, 1894.
Digitized by VjOOQ IC
240 ELECTRIC SMELTING AND REFINING.
of the Series system than in the Multiple ; but, on the other
hand, the total capital expenditure is higher in the latter case.
Kroupa states that, weighing the costs of the two processes one
against the other, the balance in favour of the Multiple system
works out at abofit 8s. 4d. per ton of copper, a difference which
is quite sufficient to account for the substitution of this system
for the series in so many of the works in which the latter had
been tried. — Translator.]
The arrangement of the new works of the Anaconda Mining
Company on the Multiple system, as above described, supports
the view that the Series system has not, in practical use, shown
superiority over the Multiple system.
TREATMENT OF IMPURE ELECTROLYTES.
From the description of the phenomena of electrolysis, given
above, it is clear that there are many substances present in the
anodes which pass into solution with the copper, and which
gradually accumulate in the electrolyte ; and it will be under-
stood that these may render a continuance of electrolysis im-
possible on account of the danger to the good qualities of the
deposited copper caused by their presence in large quantities in
the bath. But such solutions, although unfit for further use as
electrolytes, contain too large a proportion of valuable metals
(including copper) to allow of their being run away with the
waste waters of the works.
The precipitation of the copger by means of metallic iron^
although very frequently referrred to, need not be treated of here.
There is no lack of ferrous sulphate (iron vitriol) as a by-product
of metallurgical works, and in most cases, indeed, it is scarcely
saleable. There is more to be said in favour of those processes
which aim at the recovery of the copper from the solutions in
the form of copper sulphate.
In the case of works which use a closed system of circulation^
constantly passing the same liquid through the baths, it is to be
recommended that a portion (generally 20 per cent.) of the
solution in circulation should be withdrawn from time to time,
and replaced by pure copper sulphate solution. In works which
employ air as the agent for promoting circulation, and are thus
able to keep the liquid in the several tanks separate, it is
obviously an easy matter to replace the whole of the liquid in
any one tank with fresh solution from time to time as required.
There are two principal and entirely different systems of
treating these solutions in use in different works. In the one
system, the attempt is made to recover the copper sulphate
from the liquid, and in the other the removal of the impurities
is aimed at, in order that the purified electrolyte may be returned;
to the baths at the first opportunity.
Digitized by VjOOQ IC
COPPER. 241
Process in Use at the Perth Amboy Works. — The
fomier of these methods appears, from an aocoont given bj
T. XJlke,* to be in use in the works of Gaggenheim Bros, at
Perth Amboy. The solution is run into lead-lined vats, where
it is boiled with waste copper, in the presence of air and steam,
in order to neutralise the free acid present and to increase the
proportion of copper in solution. The liquid is then pumped
into crystallising pans containing suspended lead strips, on
which the copper sulphate crystallises out The mother liquor
drawn off from these pans contains practically the whole of the
arsenic and antimony originally present, together with copper
to the extent of several units per cent. This copper is next
deposited from the solution by means of sheet iron, the precipi-
tate being at first pure copper, but afterwards becoming black,
owing to the co-precipitation of arsenic, which may in the end
be present in the deposit to the extent of 60 per cent. This
impure precipitate obtained by means of iron may be either
worked up into impure copper, or may be used for the prepara-
tion of arsenic compounds, such as Scheele's green, Paris green,
and arsenious acid.
Process in Use at the Baritan Works. — At the Raritan
Copper Works the impure electrolyte is first pumped by means
of steam injectors into a building in which it is caused to trickle
through a series of so-called oxidising tanks. These are con-
structed of wood with lead linings, and, after the manner of the
Harz sulphuric acid process, are charged with granulated and
scrap copper. A current of air is directed on to this copper
as the solution trickles over it, so that the free acid may be
neutralised as far as possible. The liquid flowing from the tanks
is run into settling tanks where it is allowed to remain at rest,
after which it is concentrated to from 38° to 40"* B^. in evaporat-
ing pans, and is finally run into crystallising tanks. The
whole plant comprises for the above purpose 4 oxidising tanks,
5 settling tanks, 3 concentrating pans, and 30 crystallising
tanks. A washing apparatus and a revolving screen are arranged
for the rinsing and sizing of the crystals of copper sulphate
produced. The sieves rotate on hollow shafts, through which
hot air can be blown to dry the crystals. When it is not found
possible to obtain any more pure copper sulphate from the solu-
tion by crystallisation, the small amount of residual copper is
precipitated by means of scrap iron. In these works, about
2 per cent, of the copper introduced with the anodes is recovered
as copper sulphate, of which a portion is returned to the tanks
for the preparation of fresh solution, so that about 1 per cent,
finds its way into the market as copper sulphate.
The Smith Process. — Smith * recommends that the mother
• ZeUschriftfur EUktrochemie, 1898, vol. iv., p. 309.
t U.S. A Patent 617,996, Jan. 17, 1899.
16
Digitized by LjOOQ IC
242 £LECTRIC SMELTING AND REFIMINQ.
liquor, after crystallising out the bulk of the copper sulphate,
should be again concentrated by evaporation in lead pans, such
as are used for the concentration of sulphuric acid. During this
evaporation process the solution exposes a large surface at a
comparativelj high temperature to the action of the air, and the
iron salts are therefore readily oxidised and precipitated. When
the solution has attained a strength of 55*" B6., it is run through
cooling apparatus into settling tanks, where the remainder of
the copper sulphate crystallises out, and thence into precipitating
tanks, in which antimony and arsenic are precipitated with the
aid of thioeulphate, after the solution has been i^ain diluted.
The decomposition of the thiosulphate produces free sulphur,
which for the most part combines with the arsenic and anti-
mony, and also sulphur dioxide. It is of importance to retain
the last-named substance in the solution if the latter is to be
used again, because in electrolytes which contain sulphurous
acid, the arsenic of the anodes is converted into arsenious
add, and in this condition sinks, for the most part^ insoluble, to
the bottom with the anode>mud ; it also tends to prevent the
formation of ferric salts. For this reason the solution should be
cooled as much as possible before the thiosulphate is added to it.
After the liquid has become clear again, owing to the settling of
the sulphide of antimony and arsenic, sulphuric acid is added,
and the solution, now strongly charged with sulphurous acid, is
returned to the tanks of the refinery.
Treatment of Impure Solutions in the Chicago Copper
Bellnery. — In the Chicago Copper Refinery the practice is to
concentrate the mother liquor by evaporation, whereupon a
crystalline mixture of copper sulphate and arsenious acid crystal-
lises out. This mixture is then treated with just sufficient
water to re-dissolve the former salt, but not the acid ; and the
resulting copper sulphate is then put into use, whilst the solution
from which the original crystals had separated is treated for the
production of sulphuric acid.
According to the statement of Ulke, above referred to, l^ose
processes which aim at the purification of the electrolyte, and
the direct use of the solution again after the impurities have
been eliminated, have not given very good results in practice.
The object in most of these processes is, mainly, the conversion
of the arsenic, antimony, and iron into insoluble compounds, in
the first case by means of metastannic acid, in the second by
treatment with cuprous oxide, and in the third by oxidation
through the agency of air blown into the solutions. In the esse
of anode copper very rich in arsenic, a useful precautionary
measure is to alloy 0*1 per cent, of tin with the copper before
casting ; the tin is found to retard the formation of arseniates,
and also even to unite with the arsenic, forming anenite of tin.
In exceptional cases, where cheap water-power is obtainable
Digitized by LjOOQ IC
COPPER. 243
and where there is but a small sale for copper sulphate, Ulke *
recommends the treatment of the solution in special electrolytic
cells, in which the impurities (chiefly arsenic and antimony) are
deposited on copper cathodes, opposed to lead anodes, by means
of a current which is not strong enough to deposit iron from
the solution, but which throws down the copper, arsenic, and
antimony. Where water-power is available, as at Great Falls,
Montana, the cost of the current is not excessive, and currents
of, say, 300 amperes per sq. metre [2S amps, per sq. ft.] may be
used with advantage even in the ordinary process of depositing
copper with copper anodes. The solution thus freed from arsenic
and antimony, and rich in sulphuric acid, is made up to the
normal composition by the addition of water and copper sul-
phate, and is then used again for electrolysis, and the process
is repeated until, at last, the solution contains so much iron
that it appears to be desirable either to separate the iron salts
or to work up the solution for the recovery of copper sulphate.
The above process necessitates the precipitation of the principal
impurities. This is accomplished in covered lead-lined electro-
lytic cells, of which 12 (three groups of four cells each) are
provided for every 280 refining tanks. The cathodes are about
2 mm. thick, 900 mm. high, and 200 mm. wide [0-08 x 36 x 8 ins.].
The lead anode-plates are of the same size and are welded on to
copper supporting bars. The precipitated metals in part adhere
to the cathode, and in part fall to the bottom. The cells are
cleansed from the accumulated metallic mud once every two
months. This mud contains from 40 to 60 per cent, of copper
and is treated [metallurgically] by Bessemerising or in a re-
finery. When the cathodes are very thickly covered with
impurities, or show themselves to be otherwise unserviceable,
they are melted together and worked up into c€tke copper or
other sorts of copper, in which the presence of arsenic or anti-
mony is unobjectionable, or may even be of direct advantage
because it promotes the fasion of certain alloys. The solution,
thus purified and made up to the normal composition, is now
run into the decomposing tanks, and again used until iron salts
have accumulated in the solution to such an extent that the
cathodes become rough or darker in colour. Only in this case
is the solution treated for the recovery of the copper sulphate.
TREATMENT OP ANODE-SUMES.
In those cases in which the German cupellation process is in
use either in the same or in neighbouring works, the slime,
after screening out the larger fragments of anode in the wet
state and then drying it, may be treated direct on the hearth
by working with the rich lead which is there under treatment.
* ZeiUchiiftfiir BlektroeJiemUj 1896, voL iv., p. 313.
Digitized by LjOOQ IC
244 ELECTRIC SMELTING AND BBFININO.
According to T. Ulke,* the dried slimes are washed in the lead
in the refineries of the Pennsjlyania Lead Company. Other
works, on the other hand, make the slimes into briquettes
with lime, and then work up into rich lead by reduction. But
in the newer American works it is preferred to dissolve out
the copper, together with arsenic, antimony, and the bulk of
the other impurities by means of sulphuric acid in the presence
of air. t
Treatment of Anode-mud in the Works of the Baltimore
Eleotric Befining Company. — In the works of the Baltimore
Electric Befining Company, the slimes are treated in a lead-
lined wooden tank for from two to three hours with dilute
sulphuric acid (1 part of acid to 4 parts of water) and air, the
latter being blown into the mixture in a continuous stream by
means of a Korting's steam -jet injector. In this short period,
practically the whole of the arsenic, together with the bulk of
the other impurities, passes into solution. The liquid is then
allowed to settle, after which the clear solution is drawn off.
The residual mud, which now only contains lead sulphate,
tellurium, and a small quantity of bismuth and antimony, is
washed, dried, and fused on the hearth of a refinery furnace,
where it forms a small quantity of brownish slag containing
about 20 per cent, of lead aud 10 per cent of antimony. This
slag is drawn off, and, after solidification, is broken up in order
to recover the greater part of the enclosed silver granules, and
is then treated with lead scrap in the refinery. The lead takes
up the silver and gold still remaining in the slag, and after
enriching until it carries 60 per cent, of silver, leaves a slag
so poor that it may be smelted in a blast furnace for the
recovery of lead products. After the first slag has been with-
drawn from the molten metal obtained, as above-described, nitre
is thrown upon the surface of the metal. The slag thus pro-
duced is so rich in tellurium that if there were any sale for
that element, the refinery of this one works would yield 3 kilo-
grammes [6-6 lbs.] daily. After this treatment, the slag having
been removed, the silver is cast into plates and treated electro-
lytically for the separation of the gold. The copper contained
in the sulphuric acid solution drawn off the mud after the
steam-injection process has been applied, is precipitated by
means of scrap iron, cast into bars, and placed on the market
as crude copper.
Prooess used at the Baritan Copper Works. — At the
Baritan Copper Works, to which allusion has been made, after
the clear solution has been run out of the tank which is to be
cleaned up, the slimes are withdrawn through a plug in the
bottom of the tank and received in a movable collecting-vessel,
* Engineering and Mining Jmimal, 1896, vol. lii., p. 512.
t Cf, Wohlwill's statements, pp. 199 to 207 above.
Digitized by V^jOOQ IC
COPPER* 246
in which thej are carried to the so-called silver building. Here
they are first passed through a screen of 8 meshes to the linear
inch (3 meshes to the cm.) which retains the larger fragments
of the copper anode. The latter are washed, and the slimes
are then passed through a sieve with 60 meshes to the inch.
The residues from both sieves are washed and returned to the
melting furnace used to cast the anode-plates. The slimes
passing through the second screen are run into settling tanks,
from which the solution is presently run into receptacles placed
outside the building. The slimes are next treated with sul-
phuric acid and compressed air while heated by means of a
steam coil and agitated by paddles, until the copper is almost
completely converted into sulphate. The copper sulphate solu-
tion is conveyed to the plant in which the impure electrolyte is
worked up into copper sulphate. The slimes, now freed from
copper, are run into steam-heated drying pans, and after drying
(when they contain from 40 to 50 per cent, of silver) are treated
with lead in cupelling furnaces. After cupellation is complete,
the silver is treated by the sulphuric acid process for the separa-
tion of the gold contained in it ; and the resulting sulphate
solution is precipitated, dried, and fused according to the well-
known methods. The residual gold in the sulphate liquors is
dissolved, repredpitated, and fused to give gold in the usual
way.
Summary of the more important Industrial Conditions
obseryed in Electrolytic Copper Befining — 1. Anodes, —
Containing precious metals, otherwise consisting as far as possible
of refined copper.
2. Cathode-plcUea. — Consisting of thin plates of copper electro-
iytically deposited on lead or on greased copper plates.
3. Electrolyte, — Solution containing not less than 12 and not
more than 16 per cent, of CuSO^ -i- 5H2O [crystallised copper
sulphate] and from 5 to 10 per cent, of free sulphuric acid. It
is of great importance that the constantly-decreasing proportion
of acid and the ever-increasing proportion of copper do not vary
beyond these limits respectively throughout the whole period of
electrolysis. The cause of these alterations is to be found in the
action of air and sulphuric acid on the copper which often passes
into the slimes in considerable quantities (2 per cent, of the
anode copper), through discharge of the cuprous ions thus: —
+ + + +
Cu -f Cu = Cu + Cu.
T
A small proportion of chloride in the electrolyte is of advantage
when low current-densities are used, as it checks the formation
of cathode-growths, probably by hindering this decomposition at
the cathode.
Digitized by V^OO^ It!
246 ELECTRIC SMELTING AND REFINING.
4. Temp&nUure, — ^An increase of temperature up to 40'' C.
[104"* F.] on the average tends to improve the strength of the
copper.
5. CurrefU-conditiona, — Current-density = 40 to 150 amperes
per sq. metre [3*7 to 14*4 amperes per sq. ft.]. The proportion
of cuprous ions is greater with low current-densities, and there-
fore also the quantity of anode-slimes and of copper contained in
them, on account of the discharge on their surface of cuprous
ions (see condition 3) from the immediate neighbourhood of the
anodes. Copper may also be formed at the cathode when low
current-densities are used, owing to the incomplete discharge of
Cu, and give rise to crystalline growths through the action
described above (compare condition 3).
Potential Difference at Terminals of Bath, — 0*1 to 0*3 volt,
according to the current-density and the purity of anodes and
electrolytes used.
6. Products obtained, —
(a) Cathode copper (97 to 99 per cent, of the anode
copper.
U S^X Sffi™, *«. } '"» "■• "»^"°-
{d) Copper sulphate ) from the anode-slime and
{e) Impure arsenical copper J from pure electrolytes.
Treatment of Sulphides and Ores. — Reference will be
made in the chapter on nickel and the precious metals to the
treatment of alloys of these metals with small proportions of
copper. Although the electrolytic treatment of other copper-
bearing metallurgical or waste products or ore has not so feur led
-to any great and permanent success, the experiments undertaken
to this end are not less worthy of study than those connected
with the more fortunate process of copper refining. The study
of the treatment of sulphide products and ores must therefore
not be neglected, nor must the solution of the problem be
regarded as impossible on account of a previous want of success.
Undoubtedly the first failure met with in this direction,
beset as it is with obstacles, was that those who took the matter
in hand allowed themselves to be led away into following the
objects aimed at in electrolytic refining. Thus they attempted
the utilisation of all the arrangements that were known to be
available in refining, as types for the treatment of an anode
material that required an altogether different method of hand-
ling. Both the composition of the anodes and the residue that
was left after electrolysis, were different from those to which
metallurgists were accustomed in the treatment of metallic
anodes.
It has been shown that in the preparation of anodes for
electrolytic copper refining a metal is commonly used in which
the impurities amount to only a fraction of 1 per cent of the
Digitized by V^jOoQ It!
COPPER. 247
total metal present ; and it must be remembered that these im-
purities oonsist largely of metals, which do not take part in the
chemical reactions of the process, and which are not chemically
combined with the copper, but only alloyed with or simply
dissolved in it Then, after electrolysis has commenced, the
copper is gradually and more or less uniformly removed from
the surface of the anode, whilst a slight insoluble residue is at
first left in its place. This residue, consisting for the most part
of the precious metals, copper, cuprous oxide, lead compounds,
and the like, is very small as compared with the weight of the
copper in which it was dissolved, so that it separates in a non-
coherent form, and, becoming detached very soon from the metal
plate, collects (on account of its high specific gravity) as an
anode-slime at the bottom of the bath. Hence the anode surface
remains comparatively clean, and the small quantities of im-
purity which cling to it for a short time, offer scarcely any
hindrance to the progress of the work ; they are for the most
part conductors.
But these are conditions of so exceptionally favourable a kind
that they are but rarely met with in the whole course of metal-
lurgical practice. It is true that an alterable material has to be
dealt with in the treatment of anodes prepared from ores or
mattes, because the sulphides of which the anodes are formed
are decomposed by the current in such a manner that the metal,
leaving the sulphur behind, passes into the electrolyte, and, so
far as it is capable of precipitation under the conditions obtain-
ing, it is deposited upon the cathode. The reactions by which
this is brought about may be left for future consideration. It is
well known that in ores and mattes copper is not the only
soluble metal to be dealt with, and all previous work has shown
how important it is for continuous work that as few foreign
substauces as possible should pass with the copper into the
electrolyte. There are clearly, therefore, the most cogent
arguments against adopting the method of treatment which is
possible in the case of copper refining. It may here be re-
marked that it is sometimes desirable to make test- experiments
on a large scale, even in the face of unfavourable prospects.
Bat in this instance, costly installations have often been almost
recklessly erected, with the object of investigating disagreeable
truths that might have been discovered at a much less expense.
A continuous experiment with one, or at least with quite a few,
baths of the same dimensions as would have been used on a
large scale and requiring only about 2 to 3 H. P. would have
sufficed to show all the difficulties connected with this method
of work.
The Marohese Frooess. — The Sodetd anonima ItcUiana di
Miniere di Rame e di EleUrametallvrgia in Genoa has, at a great
sacrifice, obtained proof of the impracticability of this method of
Digitized by VjOO^ It!
248
KLECTRIC SMRLTIMO AMD REFINING.
treating copper mattes, for they erected a plant of some 125
H.P. at their works at Casarza in order to pat to a practical
test a process which was ascribed to Marchese.* According to
the account given by Badiaf the work and the installation at
this place were as follows : —
The smelting of the coarse metal (matte), of which the anodes
were made, was accomplished in the usual way ; and, at firsts a
matte containing 30 per cent, of copper, 30 per cent, of sulphur,
and 40 per cent, of iron was considered sufficiently good. The
moulds used in casting the anodes are shown in Figs. 143 to
Fig. 144«
^
^>
Fig. 145.
Mould for casting anodes of matte at Casarza.
146, which need no further description. The anodes measured
800 X 800 X 30 mm. [31^ x 31^ x IJ ins.]. Each mould was
provided with a clamp, which served to hold in its place a copper
strip that was to be inserted in the anode. This strip was used
subsequently for making the necessary electrical connection with
the dynamo during electrolysis. It may here be pointed out
that, according to a later proposal of Stolp's,{ a copper wire net
should be cast into the block of matte, in order, on the one hand,
to give more stability to the anode-plate, which is very easily
ruptured, and, on the other hand, to ensure a more uniform dis-
• German Patent 22,429, May 2, 1882. [English Patent 1,884, 1882.]
t La Ltimihre Ehctrique, 1884, vol. xiv., Nos. 40, 42, 44.
X Engineering and Mining JonmaJ, New York, 1886, ^
Digitized by LjOOQ IC
COPPER.
249
tribution of the current to the salphides. Bat even this
proposal has not been able to save the Marchese process. The
plates must be very slowly cooled after casting ; for this reason
the moulds were covered with a good heat-insulating material,
and allowed to stand during the casting and cooling opera-
tions in pits sunk in the floor of the melting-room. It was
Fig. 147. Fig. 148.
Connections of anode-strip with main conductor at Casarza.
not, of course, proposed to suspend plates of such weak and
brittle material by the copper plates inserted in them. They
were to be placed upon a wooden support provided for them in
the baths. The copper strips were, however, bent around strips
of wood in the same way as those which were also used for sus-
pension (compare Fig. 125), and were then fastened to the
conductors, which lay along the sides
of the vat, in the manner indicated in
Figs. 147 and 148.
The Cathodes, which consisted of
thin copper sheets, 700 x 700 x 0*3
mm. [27^ x 27^ x ^ ins.], were hung
in the usual way from wooden rods by
means of strips of copper, one of whicli
was carried along the upper side of the
rod direct to the main conductor, 30
mm. [1^ ins.] thick, with which it made
contact in the way described above
(Figs. 147 and 148).
The arrangement of the electrodes
in the baths is shown in the scheme sketched in Fig. 149.
The baths themselves were lead-lined wooden vats, 2,000 x 900
X 1,000 mm. deep [6 ft. 6 ins. x 3 ft x 3 ft. 3 ins.], of which
twelve were united in a group to be served by one dynamo.
The method of joining together the wooden sides and the leaden
linings of the vats at one and the same time is worthy of note.
Digitized by VjOO^ It!
Fig. 149. — Scheme of eleo-
trioal connections at
Casarza.
250
ELECTRIC SMELTING AND REFINING.
The lead sheet was screwed down with the wood instead of being
soldered therewith in the customary fashion. Figs. 150 and 151
sufficiently explain this method of construction.
The Electrolyte consisted of an acid solution of copper and
Fig. 150. Fig. 151.
Method of attaching the lead linings to the vats (Casarza).
Fig. 152. — Section of vats, showing method of circulating the
electrolyte (Casarza).
Fig. 15.3. — Plan of vats, showing method of circuUiting the
electrolyte (Uasanut).
Digitized by V^jOOQIC
COPPER.
2^1
uron BiUphates, and was obtained by roasting a portion of the ore
and extracting the roasted material by water acidulated with
sulphuric acid. The uniform circulation through the baths,
which were arranged in step form, was brought about with the
aid of leaden pipes and of wooden channels on the bottoms of
the vats. The solution flowed through a gutter to the vat which
Fig. 154. — Plan of the Casarza installation.
Fig. 155. — Croes-section of the Casarza depositing-room.
stood at the highest level (Fig. 152), and, passing through the
whole length of this vat^ overflowed into the next, as shown in
Figs. 152 and 153. The arrangement of the installation is
illustrated in the accompanying three drawings, of which Fig.
154 is the ground plan, Fig. 155 the cross-section, and Fig. 156
a general view of the interior of one of the depositing-rooms.
With anodes of such complex character, it is obvious that the.
Digitized by V^jOOQ iC
252
ELECTRIC SMELTING AND REFINING.
reactions which take place during dectrolysis must be of a very
varied character. A decomposition of the sulphides of which
the anodes are composed must precede the migration of copper,
iron, or other material capable of forming ions. And this was
supposed to have happened partly through direct oxidation and
partly through the action of ferric salts, thus, e.g. : —
CujS + 2Fej(S04)8 = 2CUSO4 + 4FeS04 + S.
Remembering the reactions which take place in the electro-
lysis of blister and refined copper, and as a result of which
many impurities gradually accumulate in the electrolyte, im-
Fig. 166. — Interior of the Casarza depositing-room.
purities that are here present in far greater quantity from
the outset, it cannot require long consideration to decide that
a considerable proportion of electrical energy must be wasted
through changes, which take place both in the solution and
upon the anode itself.
It is unnecessary here to enter upon a calculation of the
electromotive force that should be required for this process.
It will suffice to point out that after preliminary experiments,
1 volt per bath had been fixed upon as the maximum that could
be necessary, and on this assumption Marchese based his whole
estimate of the cost of the undertaking. But as the account
Digitized by VjjQO^^lt^
COPPBB.
255
(to be given shortly) of the experience gained by the Aktien-
geaeUschqft fur Bergbau^ Blei- urui ZinkhiUtenbeHeb at Stolberg
(Bheinland) will show, the expenditure of power increased very
considerably even after a short time of action.
Precipitate
of Copper
Cu S,
wfwnompnu
^remiiitmO
o a □
Rn Mater/af, Jllwkelable Products. Bye- products.
Fig. 157.--GQneral scheme of the Marcheee process.
In 1885 Marchese* published an account of a nine-day trial
that was made in that year with the experimental plant, but h&
makes no mention of this fact, ascribing the increased expendi-
ture of energy to some bad contacts that had been overlooked
by the operator. At that time Marchese still estimated a net
♦ TraitemeiU6ieclroiytique dot Mattes Cuivreusea aw Stolberg^ G^nes, 1885.
Digitized by VjOOQ IC
254 ELECTRIC SMELTING AND RRFININO.
profit of 75 per cent, upon capital outlay on plant in a factory
of this character. The scheme for the systematic treatment of
ores by the Marchese system is shown in Fig. 157.
In addition to the aboYe-mentioned favMs of this system, the
great fragility of the matte anode-plates is a serious difficulty.
Then there occurred a separation of insoluble and non-conduct-
ing material on the surface of the sulphide, and small cavities
gradually developed into deep pits during the progress of the
electrolysis. And both these circumstances caused irregular
solution of the anode, increase in the resistance, premature
crumbling of the plates, inferior extraction of the anode
material through the formation of a crust of insoluble matter,
and other actions prejudicial to the working of the process.
In place of the electromotive force which had been calculated
as sufficient for the satisfactory operation of the plants it
frequently became necessary to employ double, or even three
or four times that voltage, in order to avoid the absolute
stoppage of the work.
These last remarks have been reproduced from a paper
written by the author, in 1893,* on the direct electrolytic
treatment of ores and metallurgical products. At the same
time he gave an account of his own experiments with ore and
matte in the years 1883 to 1887. He added that these negative
results were obtained after a series of experiments lasting over
a few weeks, and conducted with the aid of a 1 H.P. engine,
and yet they established everything that the later plant of
125 H.P. at Casarza had shown.
The Stolberg Installation. — The Stolberg Ck)mpany also
determined relatively quickly upon experiments with a large
plant. Laboratory experiments liaving given promising results, a
plant was first run for several months with a 5-volt x 150-ampere
dynamo, or, in other words, with a current sufficient to deter-
mine either the conditions for, or the impracticability of, working
on a large scale. This plant was inspected by Marchese ; and
then, after doubts had been set at rest by Marchese, the erection
of a plant of sufficient capacity to deposit 10 to 20 cwts. of copper
in 24 hours was determined upon. It consisted of 56 baths, such
as have been already described, consisting of lead-lined wooden
yats, each 7 fb. 2 ins. long by 3 ft. 3 ins. wide by 3 ft. 3 ins.
deep.t Each bath contained 15 anodes and 16 cathodes, arranged
in parallel, and at a distance of 2 ins. apart.
The anodes at Stolberg were prepared from mattes of three
different grades of concentration, containing respectively 7 to
8i*per cent., 15 to 20 per cent, and about 50 per cent, of copper.
* Berg- vmd BiiitenmitnnMie Zeitwig, 1893, voL lii., pp. 251, 2S9.
t This and the following account of the reeults of the Stolberg enwri*
ments are taken from an aoooont published, with the sanction of^the
StMerger OesellBehqft, in the Zeksehriftf&r Mektroehemie, 19H, p. 50.
Digitized by VjOOQ IC
COPPER.
255
Of these, the seocmd was used directly for the production of
the anodes, the first was concentrated hj roasting and fusing
with siliceous materials, and the third was employed for the
preparation of the solutions. The composition of two separate
samples of the matte, as used for the anodes, is given in the
following table, and as the composition was variable, a third
eohimn is appended, showing the average of a large number of
analyses made at a later period than the others : —
TABLE SHOWING COMPOSITION OP ANODES USED AT
STOLBERG.
AiuilyBiaat
Stolberg.
Analysis at
Genoa.
Avenue
Compoiition.
Per eent.
Per cent.
Pbroent.
Copper (Cu),
Leaa(Pb), ....
Iroii(Fe), ....
17-20
24-78
15 to 16
23-70
12-74
14
29 18
34-23
41 to 42
Sulphur (8),. . . .
21-03
27-94
25
Sulphuric anhydride (SO*), .
0-69
...
...
Silica (SiO,)
0-88
...
Silver (Ag),. . . .
0-0623
6-656
0-05
The anodes were 80 cm. hieh, 80 cm. wide, 4 cm. thick [2 ft.
71 ins. X 2 ft. 7^ ins. x 1^ ins. J. They were made by allowing
the melted matte to flow out of the furnace into a large iron
tank from which it was removed by iron ladles and cast in iron
moulds. The moulds were sunk in the ground so that the
sulphides might cool very slowly, as otherwise the anode-plates
developed cracks, and were easily fractured. A long copper
strip 2 cm. wide and 3 mm. thick [f in. x ^ in.] was laid in the
mould before pouring, so that it might be cast into the anode,
into which it penetrated to about the centre. The free end of
die strip, outside the bath, was bent over and fastened by a
lerew clamp to the positive conductor, which consisted of a
copper rod of about 3 cm. [1*2 ins.] diameter, and so served to
make electrical connections with the matte. To avoid rupture
of the anodes under their own weighty for each weighed about
2^ cwts., they were supported upon two strips of wood placed
beneath them in the bath.
The cathodes consisted of copper plates 80 cm. by 80 cm. by
1 mm. [2 ft. 73 ins. x 2 ft. 7^ ins. x 0*04 in.], with four strips of
copper 2 cm. |t)'8 in.] wide riveted to each, in order that they
might be attached to cross strips which were laid upon transverse
wcwden slats placed above the vat. The connections of copper
strips with the negative conductor were like those adopted for
the anodes.
The electro^ was made by extracting the richest mattes
Digitized by LjOOQ IC
256 ELECTRIC SMELTING AND REFINIKO.
(those carrying about 50 per cent, of copper) with dilute sul-
phuric acid, and when ready for use contained about 27 to 28
grms. of copper and 15 grms. of iron per litre [about 4^ ozs. of
copper and about 2^ ozs. of iron per gallon]. In order to facili-
tate the constant circulation of the solution during the course
of electrolysis, the vats were arranged one above the other in
terrace form, and a pipe of 2 ins. internal diameter led from the
bottom of one vessel to the rim of that next below.
Two Siemens & Halske dynamos of the CFj-^-type, of a size
sufficient to deposit 5 cwts. of copper in twenty-four hours, were
used to produce the necessary current. These dynamos, running
at 700 and 800 revolutions, gave a current of 430 amperes x 35
volts. The current-density in each bath was about 30 amperes
per sq. metre [2*8 amperes per sq. ft.], and the electromotive
force required was at first 1 volt per bath.
The profits to be earned by the process were estimated by
Marchese in the following way: — Starting with a matte con-
taining 15 to 20 per cent, of copper, 14 per cent, of lead, and
005 per cent of silver, and expecting to recover ail these metals
(the copper by direct electrolysis, and the other two from the
insoluble anode residue by a subsequent treatment) he arrived
at the following results. In the ton of matte he reckoned : —
160 kg. of copper at Fr. 1-3 = Fr. 195
140 kg. of lead at Fr. 0*25 = Fr. 35
0-5 kg. of silver at Fr. 180 = Fr. 90
Value of 1 ton of matte «■ Fr. 320
But he was able to purchase the matte at Fr. 112*5 per ton,
because only the copper contents were taken into account, and
he thus reckoned upon obtaining at once a clear gain of Fr. 207*5
per ton of matte, or Fr. 1,383*33 per toa of copper. The interest
upon the capital locked up in the shape of copper in the baths
was thus estimated : there were 20 anodes of 125 kg. each in
every bath, and as there were 58 baths, the total weight of anode
material was 145 tons ; and this represented (in round numbers,
at Fr. 100 per ton) Fr. 14,500. But the anodes gradually gave
up their copper during the electrolysis, and he therefore took
the half of this sam as the average value of the material in the
baths throughout the operation, and so arrived at the sum of
Fr. 8,000.
Then 580 kg. of copper were deposited daily in the 58 baths.
But since three months were required to produce copper plates
of the usual marketable thickness, there remained in the baths
\^ X 580 = 26,100 kg. of copper (because here also the half of
the total copper is taken as an average) ; and the value of this is
Fr. 32,000. The percentage of the copper in the solution itself
was so small that it was regarded as negligible. The result of
this calculation is that an annual production of 210 tons neoessi-
Digitized by VjOOQ IC
COPPER. 257
tates the sinking of a capital of Fr. 40,000 in the form of copper
in the baths. Bat since the electrolytic copper recovered is
chemically pure, it is estimated as being worth Fr. 125 to 140
more per ton than is the ordinary copper of commerce. And
this, calculated on the annual outturn of 210 tons, amounts to
nearly Fr. 30,000, which is equivalent to an interest of 75 per
cent, on the capital of Fr. 40,000.
The Stolberg plant, when at first set in operation, fulfilled all
expectations ; the baths worked satisfactorily, and the separated
copper was pure. After a few days, however, the resistance of
the baths began to increase, and required, in some instances,
an electromotive force of 5 volts. The primary cause of this
increase was a dense deposit of separated sulphur at the anode,
which hindered the access of the electrolyte to the undecom-
posed sulphide within. But there was a still greater difficulty
to be encountered. The copper and iron being dissolved out
of them, the anodes lost all cohesion, so that large fragments
crumbled away and filled up the space between the anodes and
cathodes at the bottom of tlie vat. There they formed a short-
circuit for the current, because their conductivity was higher
than that of the electrolyte, and the bath remained undecom-
posed. Finally, polarisation must have added considerably to
the apparent resistance of the bath. The fact that such polarisa-
tion existed is proved by the observation that a lower E.M.F.
was required for the operation of baths that had been cut out
of the circuit for a few days. A modification of the process
suggested by the last-mentioned observation was tried, but was
soon given up again. Then, as the formation of PbOg was
considered to be the cause of the polarisation, the percentage of
copper in the anodes was increased, and that of lead diminished,
but with no better result j a change in the proportion of iron
also failed to produce any improvement. Further, the deposited
copper was found to contain antimony, bismuth, lead, iron, zinc,
and sulphur. Hence it was determined to abandon the use of
matte for anodes, in favour of lead, which would be insoluble.
A small-scale experiment was therefore tried, and as the results
were satisfSustory, a larger bath was erected. The electrolyte
employed was the same as had been used in the Marchese baths,
only it had the advantage of remaining serviceable for a longer
time, because the proportion of iron that it contained was not
being constantly increased through the action of the solution
upon the anodes. The E.M.F. required was 1*7 volts; and the
results were at first good. But in a short time the quantity
of copper deposited fell to 60 per cent, of that which should
theoretically have been obtained, and the potential rose to
2-15 volts.
Use of Depolarisers. — The cause of this defect also lay in
the polarisation of the anodes, which became coated with PbO^
Digitized by V^jOOQ IC
258 ELECTRIC SMELTING AND REFINING.
under the oxidising influence of the current, and so yielded
a counter-electromotiye force that opposed and weakened the
electrolysing current. It was then hoped that the introduction
of a reducing agent would eliminate this source of trouble.
Sulphurous add was selected for the purpose, and was led into
contact with the anodes that it might combine with the .oxygen
separated there, and so form sulphuric acid. A small experiment
was therefore made with a bath containing four lead anodes of
0*37 sq. metre [4 sq. fb.1 surfitce area and four copper-coated lead
cathodes of like dimensions. The electrolyte contained per litre
39 grms. of copper, 14*4 grms. of iron in the ferrous state,
3*9 grms. of iron in the form of ferric salts, and 9*6 grms. of free
sulphuric acid* The sulphurous acid was obtained by burning
sulphur, and, mixed with air, was injected into the bath. The
E.M.F. required was not diminished through the introduction
of the reducing agent, but the yield of copper was increased,
and the metal was purer, containing 99*984 per cent, of copper.
A large quantity of sulphuric acid was, of course, formed, and
this extracted so much soluble material from the copper mattes
that trouble was caused by the crystallisation of salts in the
baths. It should here be added that a patent t was granted for
the use of sulphurous acid as a depolarising agent in 1885. A
larger bath was then arranged on the lines of the experimental
apparatus. At first the gases from matte-calcining furnaces
were led into the baths to provide the sulphurous acid required;
but the gases were found to be too dilute, and were, therefore,
replaced by those from the muffles used in the calcining of zinc
ores. But various circumstances (among others, the offer of a
new process by Siemens & Halske) interfered with the prose-
cution of the work in this direction, which would have involved
the removal of the whole plant. The use of insoluble anodes,
and of the depolarising agent, was a great departure from the
original Marchese process, and it is to be regretted that the
experiments which were set on foot could not be carried
through; experiments, however, with other depolarisers, both
at Stolberg and in other installations, have not as yet led to
satisfactory results.
Before closing this account of the so-called Marchese process,
reference must be made in passing to the practical originator of
the method. The idea of treating copper mattes, either in the
form of plates or in the granulated condition, as anodes in an
electrolyte of sulphuric acid was clearly expressed by Andr^ in
the specification of his German patent. No. 6,048, of November
1, 1877. Andr^, however, went beyond Marchese, inasmuch as
he proposed an electrolytic separation of copper from nickel
* [This is equivalent to about 6^ ozs. of copper, 2} ess. of ferrous iron,
i DC of ferric iron, and 1^ ozs. of sulphuric acid per gallon of liquid.]
t German Patent 32,866, March 13, 1886.
Digitized by VjOOQ IC
COPPER.
259
during the treatment of the matte. Further reference to this
process will be made in the chapter on Nickel.
Body's Process. — In working the Marchese process, it was
remarked that a part of the current was used up in converting
the ferrous sulphate present in the liquors into the correspond-
ing ferric compound, and that this salt again attacked the anode
material. It was, therefore, only natural that more attention
should be given to the action produced by iron salts, with tbe
object of utilising it in some convenient way. The first ^tep in
this direction, at least in connection with electrolytic processes,
is to be found in a patent granted to Body;* but in purely
metallurgical works the utilisation of iron salts as a carrier of
oxygen had, of course, long been known. Although Body's
apparatus and process are not concerned with copper extraction
alone, but were intended for the electrolytic extraction of metals
A. ■::.._:••■■■
U
XT
A
3
1
L> :
'
s
D
s
!
1>
k^BHI^^H
Fig. 158.
Body's apparatus.
Fig. 159.
from ores in general, they may be here described as being the
forerunner of the well-known processes of Siemens k Halske
and Hoepfner.
The vessel, A (Figs. 158 and 159), is made of Portland
cement, and is painted within and without with an impermeable
material. The partition walls, S, which are also of Portland
cement, do not reach quite to the bottom of the bath ; and in the
space thus left beneath them are placed plates of some material
(like felt) that is pervious to water. The raised floor is covered
with a carbon plate, C, in connection with the positive pole of
an electrical generator ; and tbe inner surfaces of the outer walls
of the vessel are also lined with carbon, D, which, however, is
practically unnecessary. The metal plates, K, which form the
cathodes, are suspended in the space outside the partitions, S.
* U.8.A. Patent 338,150, Jan. 6, 1886*
Digitized by LjOOQIC
260 ELBCTBIG SMELTING AND REFINING.
A solution of ferric salts with sodium chloride is used as electro-
Ijte. The ore, which has been preriouslj moistened with a
similar solution, and is still saturated with it, is placed in the
inner space, J, and is here kept in constant motion during
electrolysis by means of the stirrer, R. The solution enters
through the opening, O, in the raised floor, follows the course of
the arrow marked in the figure over the partitions, S, and,
finally, after traversing the cathode compartment^ escapes
through the aperture, 0, in these outer chambers.
While the solution is thus slowly circulating the following
reactions take place : —
1. The metals contained in the ore are brought into solution
at the expense of the ferric salts in the electrolyte, which are
thus reduced to the ferrous state.
2. The dissolved metal is deposited at the cathode.
3. The chlorine which is set free at the anode peroxidises the
ferrous salts that have been produced, and any excess of chlorine
that may escape absorption in this way is able to act directly
upon the ore.
The Siemens-Halske Matte-BefLning Process. — Siemens
and Halske then went a step beyond this, and caused the
reactions between the ferric salts and the copper compounds to
take place entirely outside the electrolytic vessel. They stored
up, 80 to speak, the cmode energy in a part of the electrolyte in
order that they might utilise it outside the baths. The process
was thus described in the first German patent taken out by the
firm.*
The powdered copper pyrites is roasted at a moderate
temperature, preferably in a Gerstenhofer furnace, in such a way
that the iron is almost completely oxidised, whilst the copper is
contained in the roasted material, partly as copper sulphate
and partly as copper oxide, but mainly as cuprous sulphide.
The finely-divided material, after calcination, is treated with
the solution flowing from the electrolysing tanks. This leaching
is best performed in a series of vats, through which the liquid
flows successively in such a manner that it passes last through
the vat that was latest charged with the ore. The solution,
which is thus newly enriched with copper sulphate, and no
longer contains any ferric salt, is now returned to the
electrolytic cells, where it is first deprived of its copper, and
is then peroxidised, so that it may be used afresh to extract the
copper from another charge of ore. The process is therefore con-
tinuous, and the same solution may be used repeatedly until,
owing to the absorption of foreign metals previously contained
in the ore, it has become too impure for the process of electro-
deposition.
* German Patent 42,243, Sept. 14, 18S6. [English Patent 14,033, Nov. I»
1886.]
Digitized by VjOOQ IC
COPPER.
261
rp"
y:d/
This solution, for use in separate electrolytic cells, should be
introduced continuously nearly at the bottom of the cells which
surround the cathode-plates ; then rising to the top of these, and
depositing a part of its copper by electrolytic action on the
cathode on its way, it flows over the top rim of the membrane
into the anode compartment, through which it passes to make
its final escape from the bottom of this cell (Fig. 160). During
the passage of the electrolyte through the anode cell, the
ferrous sulphate that it contains is first converted into a basic
ferric sulphate, which in turn is changed into the normal ferric
sulphate by the absorption of sulphuric acid produced through
the electrolysis of the copper sulphate ; the higher specific
gravity of the latter salt causes it to
sink to the bottom of the vessel. The
liquid escaping from this compartment,
therefore, contains less copper than was
present in it when introduced into the
cathode cell; and it also contains neutral
ferric sulphate in solution. This solution
has now the power of converting cupric
and cuprous sulphides and copper oxide
into copper sulphate. In thus acting
upon the first-named copper compound
the ferric sulphate is reconverted into
ferrous sulphate, whilst the liberated
oxygen serves to oxidise the sulphide of
copper. The product from the roasting
of the ore at a low temperature, as above
explained, contains most of its copper in
the form of sub-sulphide ; but the iron
is present as peroxide, a substance which
is not attacked by ferric sulphate, and is
scarcely afiected by sulphuric acid. The
cuprous sulphide, however, is energeti-
cally dissolved by the ferric solution.
The chemical processes, which take place during the electrolysis,
and the leaching process are clearly shown in the following
equations : —
I. xHaS04 + 2CUSO4 + 4FeS04 = 2Cu + 2Fej(S04)3 + xU^O^
II. (a) arH^04 + CxLfi + 2Fe2(S04)3 = 2CUSO4 + 4FeS04 + S + a:H2S04
(6) CuO + HaS04 = ChiS04 + HjO
(c) 3CuO + Fe2(S04), = 3ChiS04 H-FejO,
id) CuO + 2FeS04 + HjO = CUSO4 + (FejOs + SOg) + Hj
A comparison of the equations I. and II. (a) shows that if
the ore hold all its copper in the form of cuprous sulphide, the
electrolyte, after passing through the leaching vats, will contain
exactly the same quantity of copper sulphate, ferrous sulphate,
Digitized by VjOO^ It!
■»
^
J-7
1
A
F
%
'l
K
I
D
Fi
. 160.— Siemens &
Salske electrolytic cell
for treating copper ore.
262
ELECTRIC SMELTING AND REFINING.
and free sulphuric acid as it did prior to electrolysis ; and that it
is, therefore, completely regenerated, and may be used again for
the electrolytic decomposition. But if the copper be present ia
the ore partly as copper oxide, it is evident from equations II.
(6), (c), and (d) that in this case the solution will be richer in
copper, but poorer in respect of iron and sulphuric acid than it
was before electrolysis.
It is not necessary to point out that the raw matte may be
used in lieu of the roasted material, because the copper is
present almost entirely in the form of cuprous sulphide. Ia
this case, however, iron would also be dissolved, and a complete
uniformity of the solutions in regard to copper and iron could
not be maintained. It is to be remarked that in the described
electrolytic process no polarisation occurs, and that the position
of the two electrode materials in the electro-chemical seriea
gives rise to no counter-electromotive force.
Fig. 161. — Arrangement of vats in the Siemens-Halske process.
Whilst, with matte anodes, an E.M.F. of 1*5 volts is necessary
to give the required current-density, 0-7 volt will suffice when
the above process is adopted. And, again, whilst in the former
case about one-third of the current- volume is used for other
reduction processes, and is therefore lost, by the altefrnative
method there is no loss whatever of this nature.
In order to produce the rapid circulation of liquid through the
vats, which is necessary for satisfactory work, the cells are placed
in terrace form (Fig. 161), and all the cathode compartments,
K|, Kg, K«, are connected together by siphons, A^, h^^ h^, in one
group, while the anode cells, Aj, Aj, A3, are similarly connected
in another series by the siphons, A;^, k^, k^. In order to maintain
the level of liquid in the vessels independent of the quantity of
solution added, the ends of the siphons in the lower vats are
Digitized by V^OO^ It!
COPPER.
263
turned upwards for the space, a, which is equal to the difference
in height, jS, between two consecutive vessels. dUg
Modified Siemens-Halske Process. — From a subsidiary
Fig. 162.
KJ
Fig. 163.
Fig. 164.
Siemens k Halske's electrolytic cell (1889).
patent ♦ taken out by the same firm, it must be concluded that
the use of this process led to diflSculties. The electrolytic cells
• German Patent 48,959, Jan. 3, 1889. [English Patent 3,533, Feb. 27,
1S89.]
Digitized by V^jOO^ It^
264 ELECTRIC SMELTING AND REFINING.
had preyiously been divided into two (positive and negative)
compartments by a membrane ; but it is shown in the second
patent specification that these membranes are liable to become
torn during electrolysis. The membranes have either too high
an electrical' resistance, or else they are not sufficiently durable,
for they stretch and allow the solutions to escape.
Figs. 162, 163, and 164 show an electrolyte cell in which this
evil is remedied. A flat vessel, G, made of wood or of other
suitable material, and coated with lead, is provided with a per-
forated false bottom, L, on which the anode, A, is extended.
The anode may consist, either of plates of retort carbon in direct
electrical connection with one another, or of perforated lead
plates covered with small fragments of retort carbon, or, finally,
of deeply corrugated lead plates containing perforations to allow
of the passage of the electrolyte. The horizontal anode is pro-
vided with the necessary insulated electrical connections, and is
covered with a layer of some filtering material that may serve to
prevent the escape of the solution surrounding the anodes. The
filter may be of felt or any other suitable organic or inorganic
material. The cathodes consist of the surfaces of the cylinders,
K, which are quite covered by the electrolyte, and are constantly
maintained in slow revolution by the waterproof belt, S. These
cylinders may be made of a wooden core coated with wax,
cement, or other material, and surrounded with a conducting
material, which is electrically connected in any suitable way
with the journals of the cylinders and the conductors, k.
The regenerated solution, consisting of copper- and ferrous-sul-
phate solutions, is conveyed in a continuous stream into the
liquid which is already covering the cylinders. The rotation of
the latter effects the continual mixture of the solution down to
the partition separating it from the anode compartment. The
tube, XJ, conducts the solution away from the space beneath the
filter at the same rate as the regenerated liquor is run into the
upper compartment through C ; and, there is, in consequence of
this, a constant but slow transference of liquid through the lilter
from the cathode to the anode compartment. Here the ferrous
salt is reconverted into ferric sulphate by the liberated oxygen,
and the ferric salt, having a higher specific gravity, sinks to the
bottom and is at once carried away through U, so that by
properly regulating the inflow of liquid, the strength of the
current, and the quantity of copper and iron in the solution,
the result of the process should be that the electrolyte in the
upper compartment loses some two-thirds of the coprer contained
in it^ while in the anode portion the whole of the ferrous salt is
peroxidised to the ferric state. The solution is uninterruptedly
conveyed from the anode cell to the extraction tank, and after
acting upon the ore powder, it circulates through the whole
system again.
Digitized by LjOOQ IC
COPPER.
265
According to later accounts* the anodes were afterwards
made of specially prepared homogeneous round carbon rods, a
(see Figs. 165 and 166), of which every 109 were bound together
into one group by means of a thoroughly insulated cast-lead
frame, forming a system 1,600 mm. long by 405 mm. wide [5 ft.
6 ins. X 1 ft. 4 ins.]. The connection with the main conductors is
made by means of the lead strips. V, cast on to the frames.
The electrolyte vessels are shallow wooden tanks rendered
watertight by a lining of asphalted jute. The anode system is
placed upon the floor of the tank, and the waste pipe is so
placed that it may readily conduct the solution away from the
apparatus. At a certain distance above the anode is the linen
filter, F, stretched upon wooden frames, and serving to separate
Fig. 166. — New form of anode (Siemens-Halake).
- 4
Fig. 166. — Newer form of tank (Siemens-Halske).
the bath into two compartments, one above the other. In tlie
upper or cathode compartment are placed the wooden plates, K,
which cover the whole area of the vat, and are overlaid on the
underside with thin sheet-copper to receive the electro-deposited
metal. Between these cathode-plates and the filter the cathode
solution is kept thoroughly mixed by means of a mechanical
circulating arrangement, actuated by the pulleys, R, on the edge
of the vessel.
These processes were put to a practical test in Stolberg by
the above-named company, but it is evident from Cohen*s
*Gru8onwerk-Magdeburg. Dcui Siemens' sche Kupfergeioinnungaver/ahren
aus Erzen.
Digitized by VjOOQIC
266 ELECTRIC SMELTING AND REFINING.
account that only a very imperfect apparatus was employed.
The want of durability both of the membrane separating the
anode and cathode compartments, and of the carbon anode rods,
difficulties in the clearing of the turbid solution obtained on
extracting the ore with the anode liquors, and all the derange-
ments that resulted from these troubles, caused an unforeseen
increase in the E.M.F. required (from 0*75 to 1*8 volts per bath),
and finally led to the suspension of the experiments.
The process appears from a published account given by
Siemens <fe Halske* to have been put into operation at works in
Spain and in the Tyrol, but it cannot be said to have come into
practical use there. This is the more remarkable since it is
to be inferred from the treatise above mentioned t that ample
experience could have been gained in an experimental plant at
Martinikenfeld, near Berlin. This appears, however, to have
been neglected, for, had it not been so, the great cost of the
plant laid down in Spain and in the Southern Tyrol, and
certainly experimented with at the latter place, would have
been saved, and the unpleasant record of an unsuccessful experi-
ments, such as it may now be assumed to have been, would have
been averted.
The result of all Dr. Borchers' laboratory experiments with
sulphide copper ores, whether raw or roasted, and with the
copper compounds resulting from the calcination of such ores,
is that in his opinion it is possible to extract the copper com-
pletely only by a considerable alteration of the leading arrange-
ments, so that the electrical part of the work must either be
changed or be omitted altogether. This work will soon be com-
pleted and an account of the results published.
The Hoepfher Process. — A similar fate has overtaken the
Hoepfher process, { which has, however, many points of consider-
able theoretical interest.
In the first account of his process, published shortly after his
first patents were granted, Hoepfiier described his discovery as
follows: — §
''I use electrolytic tanks, which are separated by reliable
diaphragms into anode and cathode compartments, and which
permit a through circulation from anode to anode, and from
cathode to cathode, through any number of cells placed in
series. In the anode compartments are carbon anodes, which
are incapable of electrolytic solution, and in the others are
cathodes of sheet copper. A solution of cuprous chloride in
* Jahrbuch der Mthtrockemit, 1895, vol. ii., p, 155.
t Grusonwerk - M agdeburg. Daa Siementfsche Kupfergtwinntmgtverfahrtn
aiis Krzen,
4: German Patent 53,782, March 1, 1888. [English Patent 4,626»
March 26, 1888.]
^ZeUschrififiir angeioandte Chemie, 1891, p. 160»
Digitized by LjOOQ IC
COPPER. 267
brine or in calcium chloride solution, or the like, flows past a
number of anodes in turn ; and a similar solution comes into
contact successively with any convenient number of cathodes.
Metallic copper is deposited upon the latter at the rate of
2*36 grms. [36*42 grains] per ampere per hour, that is at exactly
double the rate at which the same current can deposit the metal
from the solution of a cupric salt, such as copper sulphate.
"At the anode, if no cuprous chloride were present, free
chlorine would be liberated; and an RM.F. of 1*8 volts would
then be necessary for electrolysis. But the chlorine in this
process combines at once with the adjacent cuprous chloride,
and converts it into cupric chloride. In this way an E.M.F.,
amounting to nearly 1 volt, is produced in a direction favour-
able to the ^action in the electrolyte ; so that the electrolysis is
practically accomplished with a diflerence of potential of only
0*8 volt between the electrodes. The solution in the cathode
cells becomes weaker and weaker in respect of copper as it
passes through them successively, until finally it flows from
the last of the series nearly free from copper. Leaving the
electrolytic vats, it is collected for the above-described process
of circulation through the ore. The anode solutions retain their
copper, but no longer as cuprous chloride; for it has become
converted into cupric chloride, and the solution containing the
latter salt flows continuously from the vat."*^
*' The cupric chloride solution thus coming from the anodes is
used in the system of circulation to extract the copper and silver
from the finely crushed ores of these metals. Large leaching
vats of about 10 cb. m. [350 cub. ft.] capacity are used ; they
should be suitable for the treatment of the material with heated
solutions, and should be provided with a good stirring apparatus.
The solution acta upon copper ores according to the equation : —
CuCl, + CuS = S + CuaCla.
So that the cupric chloride, by combining with copper, becomes
reduced to cuprous chloride. It is well known that silver
sulphide, Ag^S, is readily attacked by cupric, and even by
cuprous, chloride, so that when this substance is present in
the ore, it is converted into chloride, as shown in the following
equation, and the silver chloride that is formed dissolves in the
chloride liquor : —
Ag^ + aChiClj = Cu,Cla + 2AgCl + S.
"The regenerated cuprous chloride solution is treated, as
described below, for the separation of silver, arsenic, bismuth,
and other substances, which would render the deposited copper
impure, and is then caused to flow to the anodes and cathodes.
At the former, cupric chloride is produced, whilst at the latter
the separation of copper takes place.
• Zeiiachrift/ur angetctandte Chemie, 1890, p. 622.
Digitized by V^jOOQ IC
f368 ELECTRIC SMELTING AND RGFININO.
*' The purification of the solution is most readily and practi-
cally effected by chemical means, cupric oxide or lime being
employed to remove the arsenic, antimony, or bismuth, which
4tre known to be especially prejudicial to copper, so that the
m^tal ultimately deposited may be pure. The silver is separated,
either electrolytically or chemically, before the copper. Any
small proportion of iron that may have become dissolved from
the copper ores is separated by the treatment with lime ; if this
separation were not effected, the proportion of iron would
gradually increase, and hence the solubility of the cuprous
•chloride would be reduced to about one-quarter of its former
value. Cuprous oxide is such a powerful base, that even the
oxide of zinc is precipitated before that of copper on the addition
of alkali or alkaline earths to solutions of cuprous chloride.
" The quantity of copper deposited in 24 hours in a sufficient
number of baths, with an interpolar potential difference of
0-8 volt, is equivalent to 43-9 kgs. [966 lbs.] per H.P. (= 690
ampere hours), allowing for a loss of 10 per cent. Since each
horse-power developed in large works only requires an expendi-
ture of 22 kgs. [48 '5 lbs.] of coal per diem in the boilers, it may
be estimated that in a well-managed installation, each pound of
coal burnt should produce 2 lbs. of deposited copper. Allowing
for the crushing of the ore and the work of stirring, 1 lb. of coal
should suffice to produce 1 lb. of copper from the original ore.
The extraction of copper ought, therefore, to be possible even in
countries in which the coal resources are of the most limited
character.
** The above-described process, which is shortly to be applied
in many places both in Germany and abroad will, as I [Hoepfner]
hope, quite displace the ordinary process of copper-smelting.
** According to my [Hoepfner's] calculation, a daily production
of UOOO kgs. [1 ton] of copper from a 5 per cent, pyrites would
require a capital outlay of about 123,000 marks [£6,1501, and
the daily working cost of such an installation, inclusive of
interest and amortisation charges, would be about 190 marks
tX9, 10s.]. The daily working cost of all other processes
utherto used would, under otherwise similar conditions, be
more than double this amount.
'*The causes contributing to the cheapness of the present
process may be stated as follows : —
*' 1. The greater depositing power of the current in the cuprous solution,
affording twice as much copper per hour as would be possible with a
sulphate process ; so that the cost of the electrical installation is reduced
by one hiuf.
**2. The higher efficiency of the chloride solution, which is able to
extract from the ore all the valuable metals that it contains.
'* .3. The considerable dissolving power of this solution (which may take
up as much as 150 ffrms. of copper per litre [IJ lbs. per gallon]), in con-
sequence of which the leaching plant may be of comparatively smaU size
and demands the expenditure of but little mechanical power.
Digitized by V^OO^ It!
COPPER. 26^
''Although I [Hoepfner] venture to consider the described
process as already [1891] proved to be sound, there is a possi>
bility that it may be even yet improved, so that the cost of
production may be still further reduced."
Unfortunately, the hopes which Hoepfner then expressed
have not yet been fulfilled, although, in a paper read in 1898
at the annual meeting of the Verb<md deutschen Elektrotechniker
at Frankfort-on-the-Maine, he showed the greatest confidence in
the final result of his work. The process was put to a practical
test in many experimental works — e.^., at Schwarzenberg in the
Saxon £rzgebirge, at Giessen, in Weidenau near Siegen, and
finally near Papenburg on the Ems. A special company has
been formed at Papenburg for the working of this and other
processes by Hoepfner, and to all appearances the company is.
still carrying out experiments in works built at that place.
Owners of works must certainly not be blamed if they refuse
to publish early accounts of their work, and of results which
have demanded a great sacrifice, both of time and money. It is
only possible to judge of such work by its commercial results ;
but of Hoepfner's, nothing has yet been made public, although,
according to his statement, the last-named works must have been
in operation for over four years. The difficulties in attacking
sulphide ores with cupric or ferric chloride are very great,
notwithstanding that the reaction takes place far more rapidly
than when ferric sulphate is used.
Experience has been gained in old-established works which,
for many years, have successfully conducted the leaching ojp
copper ores ; but this experience confirms the difficulties which
have given trouble with the Hoepfner process. Copper ores are
to be had which, having been attacked by natural processes,
contain the greater part of their copper in the form of carbon-
ates, and which yield this portion to dilute hydrochloric acid
with the utmost readiness; but in this case the reaction between
the chlorides remaining in the residue, after the bulk of the
copper has been dissolved and the copper sulphide in contact
with them, requires months or years for its completion. Indeed,
it is only completed in the heaps formed with these residues.
There are three means for hastening the decomposition without
previously roasting the ore : —
1. Crushing to a very fine powder ;
2. Heating during the period of leaching ; and
3. Constant movement of the slimes while being leached.
These, however, are requirements, of which the first swallows^
up a considerable proportion of the advantage to be derived
from its use, whilst the fulfilment of the last two inoreases the
difficulties of the process, and adds to the cost of constructing
the apparatus. Then, too, the further difficulty — ^by no means-
Digitized by v^jooy It!
270 ELECTRIC SMELTING AND REFINING.
inconsiderable — of clearing the solutions mast be taken into
account. Hoepfner, in the last-named paper, gives a very
favourable account of the extraction of raw Rio Tinto pyrites,
from which over nine-tenths of the 3*37 per cent, of copper
^Hginallj present was dissolved out in four hours, while only
% per cent, of the accompanying iron was dissolved with it;
but he makes no reference whatever in the paper to the
•cost of the process. It must, further, be observed that
under these conditions nearly as much iron as copper passed
into solution.
Hoep&er, at the end of his paper, says that " proof is thus
given that the electro-metallurgical recovery of pure copper
direct from its ores by the cuprous chloride process is bttsed
not only on sound scientific principles, but also more particu-
larly on technical merits;" but the author (Dr. Borchers)
considers that this expression of opinion is optimistic, and that
there is not yet sufficient proof of the practicability of the
process technically.
[The Sohwarzenberg Experiments. — The original account
given by Jensch of his experience with Hoepfher's process at
Schwarzenberg may not be available to many readers ; and as it
sets forth very clearly the details of the work a short abstract
may be given here. The ore used contained from 9*5 to 12*25
percent, of copper, and from 34*5 to 32*6 of iron; it was crushed
so finely that on an average 85 per cent, of any sample would
pass through a sieve with 200 holes to the linear inch, and 96
per cent, would pass a 100 sieve. The ore^was leached in large
revolving wooden drums, holding from 200 to 1,500 gallons each,
into which steam was admitted to hasten the reaction between
the ore and the solution, the latter consisting of the cupric
chloride from the anodes and calcium chloride. The drums gave
^considerable trouble owing to leakage, which increased with the
rise of temperature and with the growing percentage of cuprous
chloride in the solution. It was necessary to treat even the
richer ores three or four times with the leaching solution ; but
with the poorer samples, although half of the copper was removed
in the first extraction, ten or twelve were necessary to dissolve
the remaining half. It was found also that a large excess of
cupric chloride was necessary, because, at the temperature of the
reaction, magnetic pyrites and iron pyrites are both attacked by
this substance. The slimes were filter-pressed after the leaching
in order to extract as much as possible of the solution from
them. The anodes were of paraffined carbon, the cathodes of thin
sheet copper, the use of copper-coated carbon cathodes having
proved unsuccessful. The earlier difficulties with the parch-
ment diaphragms were here met with. These diaphragms were
previously described by Coehn as swelling up and becoming
very tender after a few days' use, so that when the carbons
Digitized by LjOOQ IC
COPPER.
271
disintegrated, as they did at the slightest provocation,^ the
fragments of the anode collected at the bottom of the bath and,
pressing against the parchment, produced ruptare. — Trans.]
The Coehn Prooess. — An interesting experiment of Ooehn'sf
appeared to remove one of the difficulties of the Hoepfner process
— namely, the need for the use of diaphragms, for which no
suitable material could readily be found for use in contact with
cuprous chloride on the one side and solutions containing cupric
chloride on the other.
Ooehn observed that when cuprous chloride was electrolysed
with a low current-density, the cupric chloride formed at the
anode sank to the bottom of the containing vessel in the form
of a solution, of which the specific gravity was higher than that
of the surrounding medium, and, collecting at the bottom, it
formed a layer of gradually increasing thickness. If the cathode
were so long that it dipped into this layer, copper became
dissolved from the former within the immersed area ; and the
deposition of copper could be effected most satisfactorily without
a diaphragm with the aid of the apparatus shown in Fig. 167.
Cus CI, Solution.
Permanent Level d^
the Liquid.
Fig. 167. — Coehn's single-compartment electrolyte cell.
A carbon anode. A, is used, of such length that it dips into a
collecting-trough formed at the bottom of the bath, and it is
suspended opposite a cathode of sheet copper, K, of only half its
length. The cuprous chloride liquor is admitted into the upper
part of the deep electrolysing vat, whilst the cupric chloride
solution, which streams downward from the anode, is withdrawn
* [It is claimed for the Street Girard carbons (English Patent 13,339,
1893) that they are capable of resisting disintegration, so that they may
be used in aqueous solutions, even with currents considerably exceeding
0*6 ampere per sq. in. in density. They are prepared by heating them
electrically to a temperature at which they soften and are converted into
graphite. — Translator. ]
f2kUschr%ft/ur EUktrochemie, 1895, vol. ii., p. 26.
Digitized by LjOOQ IC
272 ELECTRIC SMBLTING AND REFINING.
through a siphon from the deepest part of the trough in which
it collects. With a current-density of 20 amperes per square
metre of cathode area [1*86 amperes per square foot] the deposit
of co()per answers all requirements in respect of hoth quality
and quantity. The KM.F. required under these conditions
amounts scarcely to ^ volt.
Since the erection of Borchers' temporary electro-metallurgical
laboratory, several of his students have already, at his sugges-
tion, turned their attention to the working out of any methods
of treating troublesome ores and metallurgical products that
gave promise of ultimate success. In this way many valuable
observations have been made, which have led to the discovery
of practical processes for extraction of metals from raw materials
of the kind indicated.''^ In regard to the extraction of sulphides
by metallic chlorides rich in chlorine, their experience has
always been that the reactions take place very slowly even at
the beginning of the operation, and become even slower as the
available chlorine becomes more and more diluted, until, after a
comparatively short time, they come almost to a standstill under
the conditions which can be readily fulfilled in practical work —
viz., moderately fine crushing, moderate temperature, and not
too long a period of stirring the slimes under treatment. But
ores wluch are not available for treatment with chlorides, except
under especially favourable conditions, such as are only attain-
able in practice at great cost, present fewer difiiculties in
working, if they are treated in the form of slimes, not with
chlorides, as explained above, but with free chlorine directly
applied. Chlorine is efficacious even in very dilute solutions ;
and, as in practice the proportion of chlorine in the slimes is
kept constant by the gradual introduction of chlorine until the
attack is complete, the active material is not present in steadily
diminishing quantity as in the case of the chlorides above
referred to. Finally, chlorine acts at a much lower temperature
than do the chlorides. This, in all probability, points to the
desirability of combining the two processes by using the solutions
saturated with chlorine and the higher chlorides for the treat-
ment of the fresh ore, and completing the chlorination by means
of free chlorine after the removal of the solution first applied.
If the electrolytic method should prove to be the best for the
precipitation of copper and other valuable metals from the
solutions so obtained, there would be no difficulty in so con-
ducting the process that the chlorine necessary for the anode
treatment is yielded in the free condition instead of in the form
of cupric or ferric chloride.
Applications of Copper. — ^The uses of copper are very
* Denkachrijl der k. t, HochachuLe zu Aachen, published in connection
wi^h the Dusseldorf Exhibition, 1902. See also ZeiUchrifl fiir angewandte
Ohemie, 1902.
Digitized by VjOOQ IC
NICKEL. 273
numerous on account of the valuable properties of the metal.
Copper serves for the production of a large number of imple-
ments, apparatus, parts of machinery, and the like, both for
household and for factory use; in electrical work it finds a
special application in the form of wire. In the service of art it
is used both in coppersmith work and in electrotyping. In the
mixed-metal trades copper forms the basis of very many import-
ant alloys, such as bronze (including copper-tin, copper-tin zinc,
copper-manganese, copper-aluminium, and copper-silicon), brass
(copper-zinc), German silver (copper-nickel ziuc), <fec. For the
production of such copper compounds as, for example, copper
sulphate, cupric oxide, phosphor copper, so far as they do not
occur as by-products in metallurgical works^ metallic copper,
and the scrap from the rolling mill and the coppersmith's works
are commonly used.
CHAPTER II.
NICKEL.
Ooourrence in Nature. — Nickel is found in nature : in the
metallic condition in telluric aud meteoric iron (containing up to
20 per cent, of nickel) ; as sulphide (rarely) in nickel pyrites,
capillary pyrites or mlllerite, NiS; polydymite, Ni^S^; beyrichite,
NigS^ ; and, chiefly in combination with the sulphides of other
metals, in various forms of copper and iron pyrites, as nickel-
iron pyrites, (PeNi)S . NiS . 2FeS; horbacMte, (FeNi^gSg; cobalt-
nickel pyrites, cobalt pyrites, UnosBite, (NiCo)3S4; nickel glance,
arsenopyrite or gersdorflSte, NiAsS; antimonial nickel glance,
antimoidal nickel pyrites, nllmannite, NiSbS; as arsenide in
arsenicid nickel, kapfemickel, or niccolite, NiAs; chloanthite,
NiAsjj as antimonide in antimonial nickel or breithanptite,
NiSb j as a salt in nickel vitriol, NiSO. . 7H2O ; annabergite,
Ni.(AB0j2 • ^HgO ; emerald nickel, zaratite and tezasite, a basic
carbonate; and finally in many silicates such as gamierite,
noumeite, &c. The gamierite from New Caledonia is, at the
present time, the most important ore of nickel, since it is
obtained in large deposits almost free from copper. Now, how-
ever, that the process of W. Borchers and F. E. Giinther has
overcome the difficulty of treating nickel alloys containing
copper, it is probable that interest will soon centre again in
the extracbion of nickel from the deposits of cupriferous and
nickel-bearing magnetic pyrites, if they should prove themselves
adequate to meet the demand. . ^ ^^ ^^1^
^ Digitized by^^OOQlC
274 ELECTRIC SMELTING AND REFINING.
Among the raw materials for nickel extraction must be
reckoned, besides the above-mentioned ores, the mattes and
speise obtained in the smelting of nickel-bearing ores of cobalt^
copper, lead, bismuth, and silver, as well as the slags and
waste obtained in the concentration of these products and in
the manufacture of nickel steel.
Properties of Niokel. — Nickel (Ni*; atomic weight = 58*88 ;
specific gravity » 9) is a very clear grey, and brilliantly lustrous
metal, distinguished by its tenacity and malleability, which are
so high that, like iron, it may be rolled into sheets or drawn
into wire. In its magnetic and electrical properties it strongly
resembles iron. The melting point of nickel is about 1,400** O.
It alloys readily with most metals (copper-nickel, oopper-zinc-
nickel, German silver, iron-nickel, nickel steel). Like copper
and iron, melted nickel is capable of dissolving some of its
own compounds, as, for example, the oxide. At ordinary,
and even at comparatively high, temperatures nickel oxidises
but slightly in the air, so that the waste in the form of scale
produced in rolling, or in hot working, is much less than in the
case of iron. It combines readily with the metalloids, and its
compounds with sulphur and arsenic, as also the protoxide, play
an important part in the metallurgy of the metal. Nickel dis-
solves readily in nitric acid, and more slowly in hydrochloric or
sulphuric acid. Unless an oxidising agent co-operate with the
solvent, Ni'' compounds are always found ; and these compounds
usually result from electro-solution by electro-chemical means.
A great diversity is noticeable in the methods of treating the
various raw materials, mainly owing to the presence of copper,
cobalt, arsenic, and antimony. But electro-chemistry afibrds
here a possibility of simplifying complicated processes, and an
account will therefore be given, first, of the electro-chemical
processes applicable to the treatment of nickel, with special
references to the simplifications rendered possible by them, and
afterwards a general survey will be made of the practical pro-
cesses already in operation for the treatment of nickel ores and
other raw materials containing this metal.
It is unnecessary to refer again to the particulars given in
the earlier literature concerning the electro-chemistry of nickeL
Ruoltz,'^ in 1841, deposited nickel, like many other metals, from
cyanide solutions, and, in 1843, R. Bdttgerf described his ex-
periments on nickel-plating by galvanic means with the aid of
nickel-ammonium sulphate as electrolyte; and from that time
down to the present day many chemists and specialists have
sought to improve the processes of electro-nickeling by using
more suitable nickel salts or additions to the baths. At the
same time, the majority of these experimenters, many of whose
* Berzdiw^ Jahresbericht, 1841, vol. xxii., p. 410.
fJourval/arprakt. Ohemie, 1848. rolj^^^^t^en^^l^
NICKEL. 275
methods were pnrely empirical, ignored the possibility that the
electrolytic production of dense deposits of pure nickel may
depend less upon the nature of the nickel salt used than upon
conditions of working to which they had become accustomed in
individual cases, either by experience gained in other directions
or by accidents Under these circumstances it may be readily
understood that one observer obtained satis£Bu;tory results only
when working according to one formula, whilst another observer
only succeeded when using a different composition of bath. The
existence of so large a number of different baths for nickeling
may be explained in the same way.
Andre's Frooess. — The first account of experiments for the
electro-chemical extraction of nickel are to be found in a German
patent taken out by B, Andr^* in the year 1877. He recom-
mends the following method for the treatment of nickel mattes
or speise, or of impure compounds of nickel, cobalt and
copper : —
'* The raw material, which may be not only impure alloys, but
mattes and speise, are conductors of electricity and are, in the
form of plates if possible (or granulated), connected up to the
conducting leads by means of binding screws, so as to form
anodes suspended in a bath of dilute sulphuric acid. Copper
alone is thus deposited on the cathodes, for which carbon or
copper plates are the most suitable, but not to the exclusion of
otlier conductors. The nickel, which passes into solution at the
same time, is not deposited from an acid solution. The process
is continued until the solution and the free acid at first present
are nearly saturated with nickel. The last trace of copper is
separated by substituting a carbon anode for the matte or speise
towards the end of the treatment. In a short time the last
traces of copper are removed and the solution contains pure
nickel sulphate, slightly acid and with a little iron.
'^ (a) Pure nickel sulphate is obtained from this by adding a
little ammonia and evaporating in lead pans, air being intro-
duced during the process either through tubes or by means of
an air-blast impelled upon it. The iron is thus thrown down as
a flocculent precipitate of ferric hydroxide, which can be sepa-
rated by ladling, decanting, or filtering. The solution is then
evaporated to the crystallising point and the nickel sulphate is
put on the market in that form.
"(b) In order to obtain pure nickel from the solution, the
iron is first separated as described above, and the nickel is then
deposited from an ammoniacal bath. Plates of carbon, black-
leaded copper, or nickel, are most conveniently used as cathodes,
or other substances on which nickel will deposit may be
employed instead. If carbon or platinum, ko,, be used as
anode, polarisation will soon take place, and the efficacy of the
* German Patent 6,048, November 1, Wly^g-,^-^^^ ^^ VjOOqIc
276 ELECTRIC SMELTING AND REFINING.
current from the machine will be impaired or destroyed. Od
this account it is necessary to use anodes made of iron- or zinc-
plate, or of such form or material as will dissolve under the
action of the current. A membrane or, preferably, a double
membrane separates the anode from the cathode in order to
prevent diffusion. This object is attained still better if the
solution between the membranes is drawn off either continuously
or at intervals. The iron sulphate or zinc sulphate formed at
the anode is crystallised out and sold as such.
" The same method can be applied to the extraction of pure
nickel from siliceous nickel ores, such as those of New Caledonia
or from the Urals, after they have been brought into solution in
sulphuric or hydrochloric acid.
" (c) If it is desired to deposit copper and nickel together, the
mattes (speise, or other raw materials) are suspended as anodes
in an ammoniacal bath (ammonium sulphate) ; both metals are
deposited simultaneously on plates of carbon or black-leaded
copper; after removal by means of brushes, the deposit is washed
and receives further treatment as an alloy. Since the sulphur
present in the matte is in part converted into sulphuric acid
during the process, a small amount of ammonia, corresponding
to this, must be added from time to time. The iron is separated
out in a flocculent condition, as described under (a), together
with lead, which may be present as peroxide. When arsenic
and antimony are present, care is necessary on account of the
workmen. The copper and nickel co-deposited in the form of
powder may be almost completely separated from one another
by withdrawing the nickel by means of magnets, and for this
purpose the magnets of the machines themselves may be used."
Thus it is evident that, so far back as the year 1877, Andr6
proposed the use of copper and nickel mattes as anodes for the
electrolytic separation and recovery of metal, an invention
which is commonly ascribed to Marchese, although the first
account of the latter's experiments and patents appears to have
been published in 1882.*
Moreover, Andre in his proposals for the separation of copper
and nickel is on the right track so far as he deals with the
treatment of alloys ; for, as will be shown hereafter, copper is
readily transferred to the cathode from anodes containing copper
and nickel by the use of an acid sulphate solution, while the
nickel passes only into the solution, from which it can be
precipitated afterwards by electrolytic means, although it is true
that it cannot under any circumstances be deposited under the
conditions given by Andr^. It is impracticable, even with the
use of diaphragms, to deposit pure nickel from solutions of
nickel-ammonium sulphate with iron or zinc anodes, because
there is not sufficient difference in the electro-chemical charac-
♦ German Patent 22,429, May:?, 1882. ',,,,,, .,^
•"^Digitized by VjrOOy It:
KICKEL. 277
i^eriatics of the three metals — ^nickel, iroiip and zinc — in aqueous
solution.
Classen's Experiments. — Among the first experiments, at
this period, in the direction of creating an electro-chemical
nickel industry must be included those of Classen, aiming at the
application of electrolysis to the quantitative determination of
metals in ores and furnace products. Although it is not the
intention of the author in this work to dwell on electrolytic
methods of determining metals, it would not be right to pass
over an observation made by Classen in course of his experi-
ments on the electrolytic determination of nickel, because by its
means it was rendered possible to produce a good adhesive and
dense nickel deposit of convenient thickness, a desideratum
which had remained a secret to the electroplaters for a long
time even after the first publication of Classen's work. In the
first edition of his well-known book,* QtiatUitative Analyse durch
jElelOrolf/se, he describes the following process for the analysis of
nickelrcopper alloys : —
''The analysis of this alloy is extremely simple. A sulphuric
acid, solution of the two metals is obtained by adding sulphuric
acid to the solution of the alloy in nitric acid and then evapo-
rating. Copper is deposited electrolytically t from the resulting
solution. The solution, after siphoning ofif, is concentrated by
evaporation, the free sulphuric acid is neutralised with ammonia
or potash, and ammonium oxalate is added. The solution is
now heated, from 3 to 4 grammes of ammonium oxalate are
dissolved in it, and it is electrolysed hot."
Nothing had been said of this condition of electrolysis in
previous accounts of processes, although in a publication by
Xiinzel | it had been recommended to maintain a temperature
of 20** to 25° C. in depositing nickel with nickel-ammonium
sulphate solutions in order to avoid the crystallisation of salt
out of a solution saturated at these temperatures.
In the year 1883, in the first edition of Wahl's excellent
handbook for electroplaters, § mention is made, out of many
described, of one process in which the use of a higher tem-
perature is specified. In that case, the bath recommended is
said to give the best results at a temperature of lOO"" F., or
about 38' C.
Farmer's Process. — An apparatus described by Farmer in
1888 II had one noteworthy feature, which has proved itself
useful for other purposes also— namely, that of rotating cathodes.
* Classen, Qva/nt, Anal, aufdektrd. Wegt, Aachen, 1B82, p. 22.
t/6id., p. 12.
XAmil, Ber, tiber die Wiener Weltatmt, wm 1873. Brunswick, 1875,
ill., 1, p. 873.
§Wahr8 Oalixmoplastic ManiptdaiionSj Philadelphia and London, 1883,
p. 382.
H U.S.A. Patent 381,004, April 10, 1888.
Digitized by VjOOQ IC
278
ELECTRIC SMELTING AND REFINING.
In Figs. 168 and 169, A A are tanks, such as are oommonly
used for the electro-deposition of metals. In the upper part of
these vessels are rollers, B, made of wood, stone, or other non-
conducting substance, provided with coverings of canvas or
similar material, so arranged that the rollers may be easily
withdrawn from them. These rollers serve to maintain the
open-ended hollow cylinders, E, in continuous slow rotation.
The cylinders are made of brass or copper, and serve as cathodes ;
They are kept from lateral motion by the flanges, C C\ attached
to the rollers. The nickel plates, £', are bent into half-cylinders,
and are used as anodes. The metal arm, F, which is free to turn
Fig. 168.— Farmer's nickel depositing plant (longitudinal section).
^EZ
Fig. 169. — Farmer's nickel depositing plant (cross-section).
about an elbow joint in a vertical plane, carries a roller, D (also
made of conducting material, and mounted on c), by which the
cylinder, E, is pressed into contact with the rollers, B B, and is
connected with the negative pole of the electric generator.
During electrolysis the tanks are filled with a nickel solution
to the level shown in the figures, the double sulphate or nitrate
of nickel and ammonium being recommended for the purpose.
The current enters by the anode, E', of the apparatus on the
right, and passes through the electrolyte to the cylinder, £,
which makes one revolution in the minute, the motion being
derived from the rollers, B B, rotated by means of the pulley, P.
From the cylinder, E, the current passes through the roller, D,
Digitized by LjOOQIC
MICKBL. 279
and the arm, F, to a wire which conducts it to the next bath ;
and thence it returns to the opposite pole of the generator, either
immediately or after passing through one or more other tanks.
If the object of the apparatus were merely the alteration in
the outward shape of metals (especially of nickel), the means
would be found very expensive, and would not be likely to
displace older and well-tried methods. A combination of metal
refining with the production of articles of some special shape,
such as tubes or plates, is conceivable, but no mention is made
of such an intention, nor would it be of any service at present
in the treatment of uickel. The metallic impurities of nickel
(excepting the precious metals which are rarely present) cannot
well be separated by purely electrical means, at least, if the
nickel is to be obtained in the metallic condition in the same
operation. With the current-density that is necessary for the
deposition of nickel nearly all metallic impurities would also
be separated.
The Basse-Selye Process. — The firm of Basse & Selve"*^
have patented a process for the separation of nickel from iron,
cobalt, and zinc in a solution containing salts of all these
substances. The neutral or slightly acid aqueous solution of
iron, zinc, or cobalt, together with the nickel salt, is first mixed
with a sufficient quantity of an organic compound that is able to
prevent the precipitation of ferrous, or ferric, oxide, zinc oxide,
or cobaltous, or nickelous, oxide, by the addition of alkalies ; a
concentrated solution of caustic potash or soda is then added in
moderate excess, and the mixture is submitted to electrolysis.
With a current of 0*3 to 1*0 ampere (?per sq. decimetre)
iron, cobalt, and zinc separate at the cathode. The nickel either
remains in solution, or (especially by long-continued electrolysis)
separates in part as hydroxide, according to the concentration
of the alkaline liquid. When the solution is very alkaline and
the electric current is moderately strong (1 ampere or over), a
small proportion of the nickel separates out as black oxide on
the anode ; but this black deposit vanishes if it be left for some
time in contact with the alkaline organic liquid after the circuit
has been broken.
To obtain the nickel in the metallic condition, the alkaline
solution, from which the iron, cobalt, and zinc have been re-
moved, is mixed with sufficient ammonium carbonate to convert
all free caustic alkali into carbonate, and it is then electrolysed.
It is not necessary for the electrolytic separation of iron, cobalt,
and zinc that the nickel oxide should be dissolved in the electro-
lyte ; it is sufficient that the metals to be separated should be
entirely dissolved, and the nickel may then be present in the
form of precipitated nickelous hydroxide.
Either tartaric or citric acid, glycerine, dextrose, or other
• German Patent 64,251, Dec. 22, 1891.
Digitized by VjOOQ IC
280 BLECTBIO SMELTING AND REFINING.
organic compounds may be employed to retain the metals in
solation in presence of free alkali, but the first named affords
the most satisfactory separation.
The Strap Processes. — ^Another inventor, Jules Strap, in
his first patent,* claims to electrolyse a solution with a matte
anode ; but in a later specification f he specifies submitting the
matte first to a sulphatising roast, leaching the product, and
precipitating the iron in the resulting solution by means of chalk
and air, converting the nickel sulphate into chloride by the
addition of calcium chloride, or other chlorides, precipitating
this compound with lime, and, finally, reducing the oxide so
produced by means of carbon. It is also possible to electrolyse
the sulphate solution after adding ammonium sulphate, if the
free sulphuric acid produced at the anode is neutralised by the
addition of freshly-precipitated nickel oxide with continuous
stirring of the solution.
The Biokets Process. — ^A process, proposed by Rickets,}
for the separation of nickel and copper may from the outset be
considered as having but small prospect of success. In this
process the solutions of the salts of the two metals are to be
electrolysed after the addition of alkali-metal sulphates ; then,
while the copper is separated at the cathode, nickel is to be
precipitated to the bottom of the containing vessel in the form
of the double sulphate of iiickel and the alkali metal, which
becomes increasingly insoluble as the acidity of the bath
increases.
The Hoepfber ProoesB. — Hoepfner, in his English patent of
1893,§ describes a process for the electrolysis of nickel solutions
without entering into the question of how and in what way the
solutions are to be obtained. These are then acidified with a
weak and feebly-conductant oxygen acid (0.^., citric or phosphoric
acid), and electrolysed with the aid of insoluble anodes. The
anodes are immersed in cells containing a solution of the
chlorides of more electro -positive metals. Vertically-mounted,
rotating or oscillating plates serve as cathodes. The formation
of spongy masses is prevented by the application of movable
brushes or pads, and the electrolyte is kept in brisk motion by
means of pumps.
The anodes may be made of some metal that is partly or
wholly soluble in the solution, in place of those recommended,
but in that case some more electro-positive metal (zinc), which
will not be deposited with the nickel, must be used.
It is proposed to adapt the same method to the separation of
cobalt, zinc, lead,' tin, or copper.
* English Patent 4,396, March II, 1891.
t English Patent 16,8U0, Sept. 20, 1893.
t U.S. A. Patent 614,276, Feb. 6, 1894.
§ English Patent 13,336, July 8, 1893.
Digitized by VjOOQ IC
NICKEL. 281
In a later patenf*^ he gives some information as to the par-
ticular way adopted by him to obtain his nickel solutions.
Hoepfner, in his specification, states that it has hitherto been
held to be absurd (as it still is — Borchers) to mix pure
Caledonian ores (known to be free from copper) with copper
ores. But the process in question consists in making such a
mixture of copper ores with nickel or cobalt ores. The result
expected is that the losses as well as the smelting costs of both
the copper and the nickel ores will be lessened, and the recovery
of pure copper and nickel facilitated.
Hoepfner proceeds on the assumption that the presence of a
large percentage of copper in a nickel matte assists the electro-
lytic solution of the nickel by the cupric chloride process. f The
iron is precipitated from the solution obtained by this process
by adding alkali-metal hydroxide, or carbonate, and blowing
inair;{ the liquid is then electrolysed by the cupric chloride
process, and the solution freed from copper, and containing no
metal that is more electro-positive, is electrolysed in order to
obtain nickel or cobalt.
It could scarcely be hoped that Hoepfner's nickel process, any
more than his process for the treatment of copper, could come
into practical use. It is, of course, unnecessary to repeat here
what has been written in the chapter on copper ;§ but so far as
the treatment of nickel is concerned, the latest experience goes
to show that it is still undesirable, without some very special
reason, to smelt nickel ores free from copper with copper ores.
Hoepfner's statements and other publications || were permeated
with a strong feeling against metallurgical processes ; the goal
which he always seemed to keep before him was the produc-
tion of pure metal by an electrolytic method without previous
roasting or smelting. Yet he is obliged to encounter metal-
lurgical methods at every step. Nickel cannot be extracted
from silicate ores by means of hydrochloric acid or chlorine,
and for this reason it has not hitherto been possible to extract
the nickel even moderately completely from sulphide nickel
ores, for there is no ore known in which the nickel is present
exclusively in the form of sulphides. A by no means incon-
siderable proportion of the nickel is present as silicates in the
gangue of these ores, as is at once shown by dressing. A
matte fusion is therefore necessary in every case. But in
picturing the process of a chloridising extraction of a crude
matte poor in nickel and rich in iron, the first consideration
is the concentration of this matte by a fusion process. Then,
* English Patent 11,307, June 11, 1894.
t English Patent 4,626, March 26, 1888 ; and German Patent 63,782,
March 1, 1888.
X English Patent 22,030, Dec 16, 1891.
§ See above, pp. 266 to 270.
Jl See especially Elektrotechni^ke ZeitMchri/l, 1898, p. 732.
Digitized by VjOOQ IC
282 ELECTRIC SMELTING AND REFINING.
having reached this point, it is almost self-evident that the
next step would be the smelting of a crude metal. Then« if
the ore contained copper, this crude metal would be an alloy
of nickel and copper, and an electrolytic method could then be
applied with advantage. An alloy of this kind can well be
treated electrolytically, and with advantage as compared with
smelting processes.
The Miinzing Frooess. — Miinzing has patented a process*
which purports to use nickel mattes and speise equally with
crude nickel as anodes, and so aims at solving a problem which
has often been fruitlessly attacked before. The composition of
the electrolyte, in common with several other not unimportant
details, is not divulged. The description, however, contains
nothing which encourages the hope that the process will be
successful in practice.
Deposition of Pure Niokel and Cobalt. — The following
processf was used by Bischoff and Tiemann to produce the
metallic nickel and cobalt used by CI. Winkler in his deter-
mination of the atomic weights of these elements : —
For the production of nickel, a solution of the purest nickel
sulphate was used, containing 32*84 grms. of nickel per litre.
The electrolyte prepared with this contained : —
200 C.C. niokel sulphate solution.
30 grms. ammonium sulphate.
50 grms. ammonia (of 0*905 sp. gr.).
250 C.C. water.
Since nickel deposited upon platinum is with difficulty
detached from its surface, a highly-polished nickel plate, 9-7 cms.
long by 7*9 cms. wide, was, with advantage, used as cathode,
whilst a platinum plate was opposed to it as anode. A dynamo
was employed as generator, and a resistance was placed in the
circuit, of such strength that a fall of potential of 2*8 volts
corresponded to a current of 0*8 ampere. The current-density
was, therefore, D|qqs0'5 ampere. As soon as the nickel deposit
had attained a certain thickness it began to peel spontaneously
from the surface upon which it was depositing in thin, and
more or less curled, flakes, and there was obtained 13-13 grms.
of pure nickel in the course of 20 hours. The metal was white
and lustrous, but, as compared with cobalt, showed a distinct
tinge of yellow. No tarnish or specks of oxide were visible,
and the metal showed no loss whatever when heated to redness
in a current of dry hydrogen which had been previously puriiied
by passing over a roll of red-hot iron wire gauze. This proved
that the electro-deposited nickel had been entirely metallic in
character.
* German Patent 81,888, Aug. 15, 1894; cf. ZeUachnftfiir Elektroehemie,
1895-6, vol. ii., p. 197.
f Zeitschrift ffir anoiyjanische Chemie, 1895, vol. viii.
Digitized by VjOOQ IC
NICKEL. ZOd
For the production of cobalt, a pure sulphate of cobalt, mixed
with ammonium sulphate, was prepared by heating purple
oobaltic chloride with sulphuric acid. The aqueous solution of
this contained 11*64 grms. of cobalt per litre; and the electrolyte
was made up as follows : —
100 C.C. cobalt sulphate solution.
30 grms. ammonium sulphate.
30 grms. ammonia (of 0*905 sp. gr.).
500 c.c. water.
The cathode was a platinum plate 9 '4 cms. long by 5*9 cms.
wide, and a similar platinum plate served as anode. The
strength of current was 0*7 ampere at 3 volts, and the current-
density DjoQ = 0*6 ampere. The deposited cobalt weighed 8*133
grms., of which 7*319 grras. separated from the cathode in the
form of a coherent and fairly strong plate. The metal was
brightly lustrous on the side next to the platinum, but on the
other side it was dull and grey ; it was not tarnished, however,
and showed but little oxide. On heating in pure oxygen it
lost 0*23 per cent, in weight, so that it must have contained
0*55 per cent, of cobaltic oxide (CojOj + 2H2O) ; in other words,
0*32 per cent, of the total weight of cobalt had been deposited
as oxide.
In a second experiment the electrolyte consisted of: —
250 c.c. cobalt solution.
30 grms. ammonium sulphate.
50 grms. ammonia (sp. gr. =0*905).
250c.c. water.
A polished nickel plate, 9 cms. long by 7*6 cms. wide, was used
as cathode, and a platinum plate as anode. The current-strength
was 0*8 ampere at a pressure of 3*2 volts, and the current-density
was D^QQ a 0*6 ampere. The action was stopped after 30 hours,
and afibrded 2*9 grms. of metal, of which 2*2 grms. were easily
separated from the cathode in the form of thin and curled frag-
ments of plate. The metal so obtained was in parts perfectly
lustrous, but in many places was flecked with brilliant tarnished
spots or tinged with brown. On heating in hydrogen, it lost
0*15 per cent, in weight, which corresponds to 0*36 per cent, of
cobaltic oxide (OojOo -h 2H2O). Hence, 0*21 per cent, of the
whole of the cobalt had been deposited as oxide. The deter-
minations of oxide made in this way, however, are likely to be
a little too high, because the deposited metal contained traces of
ammonium salts, even after very thorough washing ; and these
were volatilised on heating in hydrogen, afibrding a slight
brownish ring of deposit in the cooler part of the tube. Alter
heating in hydrogen, the cobalt had a uniformly metallic appear-
ance, and in parts formed plates with a beautiful lustre. Its
colour, as compared with nickel, was distinctly bluish-white,
like that of zinc.
Digitized by VjOOQ IC
284 ELECTRIC SMELTING AND REFINING.
The Heibling Prooess. — A process proposed by Heibling*
to produce ferro-nickel, in the electric furnace, by the redaction
of mixtures of the oxides of nickel and iron, and at the same
time to obtain calcium carbide, by the simultaneous reduction
of calcium oxide or carhonate with an excess of carbon, cannot
be said to have any prospect of success in this connection. The
production of ferro-nickel does not demand the high temperature
which is -necessary for calcium carbide. Moreover, the fluxing
of the silica, which is contained in nearly all the ores and
furnace- prod acts of nickel, would be almost impossible under
the conditions necessary fur the production of calcium carbide.
Indeed, a patent of Bathenau's,! which is in actual operatit)n,
shows that the silicides of iron and nickel would be thus
yielded, in place of the metals which, according to Heibling,
would be expected.
The lie Verrier Process. — Le "VerrierJ proposes to separate
the nickel electrolytically as follows, from an alloy of nickel
with iron which he would obtain by the furnace-reduction of
New Caledonian (silicate) ores : — It is well known that under
the conditions obtaining in the electro-deposition of nickel, iron
and nickel dissolve with nearly equal rapidity at the anode.
The inventor proposes entirely to prevent the passage of iron
to the cathode, by using an electrolyte consisting of a dilute
solution of nickel-ammonium chloride, to which sodium chloride
has been added, maintaining a weak basic reaction in the solu-
tion, and aiding the peroxidation of the iron passing into the
solution by the addition, from time to time, of a small quantity
of calcium hypochlorite, or, if only a little iron be present, by
blowing air through the liquid. It is claimed that, under these
conditions, the iron is completely precipitated from the solution
in the form of hydroxide. Nickel-ammonium sulphate may be
used as electrolyte if sodium hypochlorite be used as an oxidising
agent in place of chloride of lime.
The Kugel ProcesB. — Kugel§ recommends the use of weakly
acid hot electrolytes, consisting of the sulphates of nickel and
magnesium, with anodes of nickel mattes or of insoluble materials.
The solution may consist of 800 grms. of nickel sulphate and a
like quantity of magnesium sulphate per litre, at a temperature
of90°C.
The Frasch Process. — The Frasch process, || which has been
tried in a Canadian works, can scarcely be expected to promise
more success than did the Hoepfner process. The matte is
roughly crushed and packed in vessels the bottoms of which are
* English Patent 18,407, Oct. 3, 1896.
t Compare the section on CarbuU^ and sUicidts.
t English Patent 5,781, 1898; German Patent 112,890.
§ German Patent 117,054.
ll Engineering and Mining Jaum,, 1900, vol. Ixx., p. 272.
Digitized by VjOOQ IC
NIGKKL. 285
covered with copper; this layer is covered with a stratum of
sand, and a strong solution of salt is ran into the lower portion
of the vessel, up to the level of the sand diaphragm thus pro-
duced. The vessel is finally filled with water or a weak solution
of caustic soda. The copper plates, with the matte resting upon
them, form the anode, whilst the cathodes are immersed in the
weak solution of caustic soda. On passing the current, caustic
soda forms in the upper part of the vessel, whilst it is intended
that the chlorine shall form chlorides with the metals contained
in the matte. The chloride solution thus formed is then to be
treated either electrolytically or by chemical means, or it may,
after a portion of the metals present (especially the copper) have
been precipitated electrolytically, be applied to the treatment of
matte by reason of the copper chloride contained in it. After
leaching, the matte is taken out, and is then, after washing, used
again, and it is only necessary to remove the residue of the
leached material.
The matte used contained, per cent., copper = 31*8, nickel =
14*8, iron = 25, and sulphur — 24. The solution first removed
from the anode cell contained, per litre, copper = 50 grms.^
nickel » 43 grms., and iron = 26*6 grms. After depositing a^
portion of the copper and passing the solution again through the
anode cell, a liquid was obtained which contained, per litre,
copper =: 50 grms., nickel = 60 grms., and iron = 36 grms. By
repeating these operations the inventor anticipated that he would
obtain solutions of any desired composition.
Separation of Cobalt. — Reference must here be made to
certain proposals for the electro-chemical separation of cobalt as
oxide, since they are of practical use in the purification of nickel
solutions.
The Vortmann Process. — Yortmann^ gives his method of
separating nickel and cobalt in the idea (which, however, cannot
be considered as acceptable under all circumstances) that the salts
of cobalt, like those of nickel, when in aqueous solution in the
absence of alkali-metal sulphates (or of other neutral salts of
the alkali metals), are decomposed by the electric current
in such a way that the hydroxide, or basic salts of cobalt
or nickel protoxides, separate out at the cathode. If, under
these circumstances, the direction of the current is, after a
time, reversed, the nickel oxide dissolves, while the cobaltous
oxide becomes oxidised to cobaltic hydroxide. Hence if, from
time to time, the direction of the current is reversed, the
whole of the cobalt should ultimately be precipitated as brown-
black hydroxide, while the nickel remains in solution. The
oxidation of the cobaltous oxide formed at the cathode takes
place more rapidly if the solution contains a small proportion
of a chloride— as, for example, about 1 per cent, of common
• German Patent 78,236, May 10, 18W. . ....... .^
Jigitized by V^jOO^ It!
286 ELEcraiG shblting and befinino.
salt. In this case an oft-repeated reversal of the current is
unnecessary, as the cobaltous oxide is rapidly converted into
cobaltio oxide by the action of the small quantity of chlorine
or hypochlorous acid distributed through the solution. Warming
the solution gently during electrolysis facilitates the separation
of the cobalt in either case. Any traces of nickel oxide precipi-
tated with the cobalt may be re-dissolved by heating the solution
for a short time to 60'' or 70" 0. after the precipitation of the
cobalt is finished and the current is no longer passing. The
nickel solution filtered from the cobaltic oxide should be free
from cobalt^ or should contain, at most, only a trace of that
metal.
An account*^ given by Ooehn and Salamon, however, shows
that Yortmann's assumption is not quite correct. The proposal
there made was to separate the cobalt electrolytically from the
nickel by depositing it at the anode in the form of peroxide, but
they were compelled to prevent the deposition of nickel and
cobalt at the cathode by the addition of a copper salt to the
solutten. « Interesting as this observation was from the scientific
standpoint, it appears that the expectation of the authors that
the process might be turned to useful account industrially does
not appear to have been fulfilled, for, in a later patent,t the
inventors recommend the precipitation of the cobalt by means of
persulphates, and so return to the old method of cobalt-nickel
separation, according to which cobalt and iron are successively
precipitated from neutral solutions by means of oxidising agents.
In industrial work the agent commonly used is chloride of lime,
or chromates or peroxides are employed when this material must
be avoided, in order to prevent the contamination of the liquid
by chlorides.
nike's Experiments. — T. Ulke,! in 1897, gave a noteworthy
account of the methods at that time used for refining crude
nickel, and of a process of his own for the treatment of nickel
mattes. Nickel oxide was in the first place obtained by reduc-
tion by tops-and-bottoms smelting and chloridising roasting from
Canadian pyrrhotite at the Orford mine at Constable's Hook,
New Jersey. This oxide was then reduced to a crude metal
containing, per cent. — Nickel, 95 to 96 ; copper, 0*2 to 0*6 ;
iron, 0-75 ; silicon, 0*25 ; carbon, 0*45 ; and sulphur (intention-
ally added to facilitate pouring), 3; with about 15 grms. of
platinum per ton. The metal was cast into anode-plates and
converted into pure nickel by electrolytic means (Ulke believes
that cyanide baths were used) in the works of the Balbach
Smelting and Refining Company at Newark. The refined nickel
* Zeitachriftfur EUktrochemie, 189ft, vol. iv., p. 501 ; and German Patent
102,370, March 4, 1898 ; English Patent 9,153, April 20, 189&
t German Patent 10,615, Feb. 5, 1899.
X Engineering and Mining Joum,, 1897, vol. hriii., p. 112.
Digitized by LjOOQ IC
NICKEL. 287
SO obtained contained, per cent. — Nickel, 99*5 to 99*7 j copper,
O'l to 0*2 ; arsenic, 0*03; sulphur, 0*02; iron, 0*1; and traces of
platinum.
For the direct treatment of nickel -copper mattes Ulke recom-
mends the following method : — Anodescastfrom matte (containing
on an average, per cent., copper, 43*4; nickel, 40; iron, 0*3;
and sulphur, 13*8 ; with 210 grms. of silver, 3 to 6 grms. gold,
and 15 grms. platinum per ton) are electrolysed with copper
cathodes in an acid solution of sulphates, such as may be obtained
by the leaching of roasted mattes. If a gentle circulation of
solution be ensured and a moderate current-density be employed,
copper alone will be deposited out of all the metals passing into
the solution. When the nickel has become highly concentrated
in the solution the copper remaining in the liquid is removed,
either by means of sodium sulphide or by allowing it to pass
through a filter of nickel mattes. The small quantity of iron
present is removed by one of the older methods or by the
newer process of Whitehead, with the aid of freshly-precipitated
nickel hydrate. From the residual solution there may be
obtained at will either (1) nickel sulphate, by concentration ;
(2) nickel oxide, by precipitation; or (3) metallic nickel, by
electrolysis.
Forster's Experiments. — Forster's experiments, made in
the year 1897,* contirmed the observations of Classen and
others that the most important condition for the precipitation of
a dense malleable nickel is the maintenance of a temperature
higher than the normal — namely, from 50** to 90° G. AH other
conditions, such as current-density, concentration of solution,
and the like, fall into the background in comparison with it,
although they have been regarded as of special importance in
electro-plating, and in many cases have been the subjects of
invention. Sulphate solutions, as Bottger showed in 1843, have
proved themselves especially suitable as electrolytes. Forster,
who has made comparative experiments, chiefly with sulphate
and chloride solutions, • finds greater difficulty in obtaining a
good nickel deposit with the latter. He found that while
with sulphate solutions, containing from 30 to 100 grms. per
litre, he could ensure a good deposit by using current-
densities of from 50 to 300 amperes, with chlorides the
solutions had to contain from 50 to 120 grms. of nickel
per litre, and a current-density of from 70 to 300 amperes
was necessary. Again, whereas in the case of sulphate solu-
tions but little care is required to maintain an approximately
neutral reaction with chloride electrolytes, at the temperature
most suitable for deposition, the solution of the nickel at the
anode takes place more quickly than corresponds to the current
applied, so that a deposit of basic salts occurs at that point,
♦ ZeiUckriftfUr BUktrochtmie, 1897-8, vol. iv., p. 160. ,
Digitized by VotOOQIC
288 ELECTRIC SMELTING AND REFINING.
unless care is taken to see that the liquid always has an acid
reaction. Forster found that an acidity corresponding to 2*5
grms. HOI per litre was necessary, and that in order to keep
this constant, hydrochloric acid must be regularly added at the
rate of 0*05 to 0*1 grm. HOI for each gramme of nickel deposited.
Forster found that when he used the lower current-densities and
the more dilute solutions above-named, he obtained dull-grey
rough deposits ; but this is also the experience of the electro-
plater, who therefore makes a point of treating objects with a
higher current-density when first placing them in a nickel-bath,
>ind then carries on the process at a lower current-density, often
of only 30 amperes [per sq. metre]. All the above figures relate
to the use of metallic (crude nickel) anodes. The drop of
potential at current-densities of 150 amperes is only 1 volt, and
at densities of 200 amperes per sq. metre about 1 '3 volts.
Separation of Nickel-Copper Alloys. — The following
account was supplied by Dr. E. Wohlwill for the German
edition of this work : —
" With alloys of this kind, the alteration which constantly
takes place in the electrolyte is a direct result of the process
employed. The separation of the two metals depends in this
case upon the fact that both are dissolved, but only copper is
separated in the pure state at the cathode ; but since the quan-
tity of copper deposited is equivalent to the sum of the quantities
of the two metals dissolved at the anode, it follows that for
every pound of nickel passing into the solution 1*08 pounds of
the copper contained in the solution is separated out at the
cathode, and no copper passes into tlie solution to take its place.
[This is, of course, independent of the anode copper which passes
into the solution, and is deposited in equal quantity at the
cathode.] Hence, as the quantity of nickel present in the solu-
tion must constantly increase, there must, in a given time,
arrive a moment after which the amount of copper present will
be so greatly reduced that it is impossible to obtain a coherent
metallic deposit, and at last only a mixture of copper and
hydrogen will be deposited.
"When this period has been reached, nothing further is
gained by the introduction of copper into the solution from the
anode, for the amount so dissolved must always be leas than
would correspond to the amount of the copper to be precipitated.
If, therefore, all the copper dissolved at the anode is to be
recovered at the cathode in a useful form — t.e., as pure coherent
metal — it is necessary that copper sulphate shall \}e regularly
added in quantity equal to that of the copper deposited as the
equivalent of the nickel dissolved at the anode. When, there-
fore, the separation of copper and nickel is to be regarded as a
regular industry, and not merely to be applied incidentally to
the treatment of a small quantity of material, the process must
Digitized by V^jOO^ It!
NICKEL. 289
be associated with the simultaneous recovery of copper from its
sulphate. The extent of the consumption of copper sulphate to
be used is determined by the proportion of nickel in the anode.
For example, if an alloy of 90 per cent, of copper with 10 per
cent, of nickel is to be treated, and supposing that the proportion
of copper sulphate in the solution is to be constant, for every
100 lbs. of alloy which passes into solution at the anode,
10 X -^ = 10*8 lbs. of copper must be added in the form of
sulphate, and therefore 42*4 lbs. of copper sulphate must be
introduced into the bath. The 10 lbs. of nickel passing into the
solution corresponds to 10 x 4*78 = 47*8 lbs. of nickel sulphate
(NiS04 . 711^0). A comparatively large proportion of nickel
sulphate may, however, be present in the solution without
prejudicing the purity of the copper deposited, provided that the
quantity of copper present is kept constant. Hence the opera-
tion may be continued until the gradually-increasing proportion
of nickel is sufficiently great to allow of the further treatment of
the solution. It is advisable that the concentration of the
nickel sulphate should not be allowed to exceed about 200
kilogrammes per cb. metre [2 lbs. per gallon], since at higher
concentrations portions of the solution are retained by the
200
cathode copper. From this it follows tiiat 100 x j=r =* 418
kilogrammes of the alloy above-mentioned can be treated per
1 cb. metre of bath [4*18 lbs. per gallon] up to the time when
the solution needs to be renewed. Simultaneously 42*4 x 4*18
= 177*2 kilogrammes of copper sulphate are used per cb. metre
[1*77 lbs. per gallon], so that in treating an alloy of copper with
10 per cent, of nickel, about one-ninth part of the electrolytically
deposited copper obtained will be derived from the decomposi-
tion of the copper sulphate added. It is worthy of note that
the recovery of copper from copper sulphate by the process just
described is eJQTected with a smaller expenditure of power, and
therefore at a lower cost, than is the case with ordinary copper
sulphate solution (unmixed with nickel) using insoluble anodes.
The electromotive force required for the decomposition of the
solution is in the latter case, at the beginning of the operation,
about six times as great as it is when nickel-copper or similar
alloys are used as anodes. Hence the production of copper from
copper sulphate may be usefully effected in combination with
the treatment of these alloys in cases in which the direct electro-
lytic decomposition into copper and sulphuric acid would be
impracticable commercially.
^'In order to obtain the nickel sulphate from the mixed
solution obtained by electrolysis, the addition of copper sulphate
to the bath is not continued as long as would be required accord-
ing to the calculation given above. On the contrary, as soluble
19
Digitized by V^OO^ It!
290 ELECTRIC SMELTING AKD REFINING.
anodes are used, the addition of the copper sulphate to replace
the copper deposited is stopped as soon as the maximum percen-
tage of nickel sulphate is approached, and the reduction in the
proportion of copper in the hath, is continued up to the limit at
which a good useful deposit is obtained. When this point is
nearly reached — after a reduction of current-density towards the
end — ^the cathodes are replaced by lead or copper plates on which
considerable quantities of less pure and less cohesive copper are
deposited, and, finally, insoluble plates are substituted for the
soluble anodes. Lead plates may be used for this purpose,
provided that the composition of the electrolyte is such that the
possibility of the formation of lead peroxide is excluded. The
writer [Wohlwill] has observed that this is the case in working
with nickel-copper alloys, if the alloy contain even a relatively
small quantity of iron. The solution of ferrous sulphate obtained
in such a case protects the lead so completely that the same
plates may serve as anodes for the purpose for an unlimited
period. In the complete absence of iron from the solution the
lead anodes are rapidly destroyed unless the water used is free
from chlorides — a condition which is rarely, if ever, to be met
with in practice. Each increase in the amount of chlorides
present in the electrolyte favours the formation of lead peroxide,
and, therefore, the destruction of the lead anodes to an extra-
ordinary extent; but even in the presence of chlorides, an
addition of ferrous sulphate in no very considerable quantity
provides the necessary protection.
" The bulk of the copper may be removed from the solution,
gas being evolved at both electrodes and the potential difference
altering accordingly. The solution, now containing iron, but
only a little copper, is evaporated to the crystallising point, the
mixed crystals of copper and iron compounds are separated by
chemical means, and finally pure nickel sulphate, or, after the
addition of ammonium sulphate, nickel-ammonium sulphate, is
obtained by repeated crystallisation."
Beview of Frooesses from Industrial Standpoint. — It
will have been seen from the foregoing account that the most
determined attempts have been made to apply electro-chemical
methods of treatment to almost all products, whether of nature'
or of art, which contain any appreciable quantity of nickel;
indeed, even the impossible has not been left untried.
As the result of the work that has been done it may now be
said that
(1) The direct electro-chemical treatment of the ore is imprac-
ticable, for no known nickel-bearing ore is rich and pure enough
to enable it to be treated, either as anode or otherwise, in an
electrolytic cell or in an electric furnace.
(2) The treatment of the ore by a combined system of leaching
and electrolysis is practically out of the question. Silicates are
Digitized by V^jOO^ It!
NICKEL. 291
not sufficiently soluble in the solvents available ; the sulphide
and arsenical ores contain so little nickel, and at the same time
carry so many constituents which not only consume solvent
but in many cases are valueless, that the cost of dissolving the
ore is increased beyond that which is permissible in view of
the amount of nickel present. Moreover, it must be added that,
as mentioned in speaking of the Hoepfner process, a not incon-
siderable proportion of the nickel in sulphide ores is present as
silicate in the gangue-stuff, and is therefore almost completely
withdrawn from the action of the solvent. With all the ores at
present known, it has been found that the first step towards the
recovery of the nickel is a process of fusion for matte or speise.
No leaching process has yet succeeded in extracting more than
50 per cent, of the nickel contained in sulphide ores, even though
the ore has been crushed as fine as it is practically possible to
reduce it. This is the less surprising when it is remembered that
even crude matte, which, in comparison with ore, contains but
little nickel in the form of silicate, has not been found to yield
much more.
Treatment of Gonoentrated Mattes. — ^Although many pro-
positions have been put forward, there are no accounts available
of any results obtained in practice in the treatment of concen-
trated mattes. If further experiments in this direction were to
be tried it would be desirable in the first place to determine the
limits economically attainable in the enrichment of the sulphides.
It is possible, by blowing [in the converter], to obtain a concen-
trated matte in which the proportion of nickel is greater than
corresponds to the formula NiS. TheNiS contains about 65 per
cent, of nickel ; but by Bessemerising it is possible to obtain
mattes containing over 75 per cent., and it is hoped that it may
even be concentrated up to 95 per cent. Ni, but in that case the
mattes would have to be regarded as consisting of the following
mixtures : —
75 per cent. Ni = an alloy of 71 per cent. NiS + 29 per cent. Ni.*
80 ,, »» = If 57 ,f ,t + 43 „ ,,
OU ,) f, = yf 4«> ,, |, + 07 }} If
90 „ »» = It 28 ft ,, + 72 ,1 „
9D >f »> * » 14 ,, ,, + oo )) ,,
But it is impossible at present to predict what would be the loss
in the slags or what would be the cost of producing these high
concentrations.
For materials containing copper the cost of Bessemerising does
not appear to be prohibitive. According t9^Ulke,t at the works
of the Canadian Copper Company the cost of concentrating by
the Bessemerising process a matte containing 40 per cent, of
* Stated in round numbers per cent.
t Engineering and Mining Joum,f 1897, vol.lxiv., p. 9.
Digitized by V^jOOQ IC
292 ELECTRIC SMELTINO AND REFINING.
copper and nickel, conjointly, to a matte with 43 to 44 per cent,
of copper and 40 per cent, of nickel (84 per cent, in all), works
out at from 50 to 60 marks {£2 lOs. to £3) per ton ; whilst the
concentration in the same way of a poorer matte carrying 30 per
cent Ni + Cu to one averaging 80 per cent. Ni + Gu is stated by
Vogt* to be only 8 marks (88.) per ton. The figure quoted by
Vogt is, however, without doubt, too low; 8 marks would
scarcely cover the cost of the blast and the power necessary for
the converter.
The question of concentrating more highly and treating eleo>
troly tically nickel matte containing copper need not be considered;
the process is only suitable for non-cupriferous mattes. For
such material it is only necessary to determine whether, and at
what degree of concentration, a given matte can be worked more
advantageously by electrolysis or by the well-known reduction
processes. So far as -experience is at present available, the latter
process must still be considered as the only practicable method.
It is true that the bulk of the iron can be removed from
cupriferous matte by Bessemerising, and the matte at the works
of the Canadian Oopper Oompany contains, after blowing, only
0'3 per cent, of iron ; but the further reduction in the quantity
of iron by prolonged blowing must necessarily lead to the slagging
off of a large quantity of nickel, and it is open to question
whether a material practically free from iron could be obtained
within the limits of concentration attainable in actual practioe.
But if it is found necessary to permit the presence of a certain
quantity of iron, it is certainly needless to go so far as has
hitherto been deemed requisite in the concentration of nickel-
copper mattes. In fact, a process described by Borchersf has
shown that the treatment of matte containing quite a large pro-
portion of iron is practicable with good results. From all that
has gone before, it is evident that the electro-chemical separation
of the metals contained in the matte does not require the use of
the latter as anode or as a material to be leached. The matte
should rather, after dead-roasting, be smelted to a nickel-copper-
iron alloy, by which means the chief difficulties in the way of
treating the matte by electro-chemical processes direct are
removed. The matte selected as the subject of certain experi-
ments contained from 60 to 70 per cent, of metal, of which from
15 to 20 per cent, was iron. After roasting and then reducing,
an alloy was obtained in which the metals, iron, copper, and
nickel, were present in the proportion of 1 : 1 : 2 by weight.
Such an alloy, when used as anode in an acid solution of copper
sulphate, yields a solution of nickel sulphate and ferric sulphate,
whilst the copper is deposited at the cathode. Andr^ and
Wohlwill, as already explained, had attained this point. It is,
♦ Slahl und Eieen, 1894, p. 23.
t Jdhresbei-icht der Mlekti-ochemiet 1897, vol. iv., p. 305.
Digitized by V^jOOQ IC
NICKEL 293
therefore, only necessary to add that electrolysis gave a good
deposit of copper until the proportion of copper in the solution
had been reduced to less than 1 per cent. Wohlwill has ex-
plained the cause of this natural reduction in the amount of
copper present, and has recommended the addition of copper
sulphate, in quantity proportional to the nickel (and in this case
to the iron in addition), entering into solution. The current-
density varied between 1 50 and 200 am])eres per sq. metre, and
the potential difference between 0*5 and 1*2 volts, according to
the quantity of copper in the electrolyte.
When the electrolyte has become saturated with the sulphates
of iron and nickel, the residue of copper present may be most
conveniently removed by means of iron scrap. The method
recommended by Andr6 and many others of precipitating the
copper electrolytically with the aid of insoluble anodes, is not
practically suitable with solutions of mixtures of salts such as
are here obtained. As the proportion of copper present becomes
less the E.M.F. required increases, and the current-efficiency
diminishes even if the current-density be reduced to a corre-
sponding extent.
Borohers' Frooess for the Separation of Iron and Nickel.
— The difficulties in the separation of iron and nickel sulphates
have undoubtedly been underestimated hitherto. The separa-
tion is, however, an easy matter by means of a process devised
by Borchers.*^ By this process ammonium sulphate is added to
the solution in quantity equivalent to the nickel sulphate
present, and the liquid is carefully crystallised, taking care that
the concentration of the solution is always maintained below
the crystallising point of a ferrous sulphate solution. A nickel-
ammonium sulphate solution crystallises best at a density
(measured at the boiling point) of about IS"" Be. (» 1*143 sp.
gr.), whilst an iron sulphate or ferrous-ammonium sulphate
solution only begins to crystallise at 31° Be. (1275 sp. gr.).
The concentration is maintained either in the neighbourhood of
18*" Be. or is allowed to exceed this, according as the amount of
nickel salt is great or small ; but it should never be allowed to
approach within 5"* B^. of the specific gravity at which the iron
salt crystallises. When the proportion of iron present is very
large, the separation of nickel and iron by the first crystallisa-
tion will be anything but satisfactory; but there has never been
any metallurgical difficulty in reducing the quantity of iron by
transference to the slag in the preliminary fusion -refinery
process, even down to a fraction of 1 per cent. On the other
hand, it was shown by the experiments, in which the success of
the method was proved to be possible, that solutions gave satis-
factory results in which the proportion of nickel was equal to
that of iron. The crystals, after washing, may be again dissolved
* Jahttibericht der ElekirochemUt 1897, vol. iv., p. 306.
Digitized by V^jOOQ IC
294: ELECTRIC SMELTING AND REFINING.
and freed from any trace of iron (introduced from mother-liquor
clinging to them) by precipitation with chromates or per-
Bulphates. The nickel-ammonium sulphate solution may then
be used as the electrolyte. Experiments have, however, been
made by E. F. Gunther in Borchers' laboratoiy with the object
of introducing this process into copper-nickel works ; and these
experiments showed that it was not possible to obtain quite
sufficient concentration for the purpose of electrolysis if the
double salt was used. It was, therefore, necessary to convert
the double into the single salt. It is well known that manufac-
turers of ammonium salts prepare ammonium carbonate from
the sulphate by the addition of soda. It was, however, undesir-
able to use soda directly in these experiments because the
ammonium sulphate in the double nickel salt would have been
replaced by sodium sulphate. But, since it is not desirable to
precipitate the whole of the nickel electrolytically from its
solutions, the residue of nickel in the electrolytically im-
poverished solution is precipitated as carbonate by the addition
of sodium carbonate, and the nickel carbonate so obtained is
applied to the decomposition of the nickel-ammonium sulphate,
which is by its means converted into nickel sulphate and
ammonium carbonate. The apparatus to be used would be the
same as for the manufacture of ammonium carbonate. The
absorption of iron by the solution cannot be entirely prevented
during the process, so that it is necessary to remove any iron
that may be present when the salts are again brought into
solution, before transferring the liquid to the electrolytic cells.
The removal may be effected by the methods mentioned above.
Utilisation of Anode Beaotions. — The author, in the
previous editions of this work, has frequently stated that, in
all cases of electrolytic deposition, if no suitable material is
present which will pass into solution at the anode or be sepa-
rated from the electrolyte in useful form during the process,
care must be taken that there is some practical return for the
work that is there expended. Bontschewski "* had shown by
brilliantly successful experiments in Borchers' laboratory that
this was possible in the case of zinc electrolysis, as, for example,
by the production of lead peroxide without the aid of dia-
phragms ; and E. F. GUnther has experimented with a view to
producing various lead pigments, such as white lead, chrome
yellow, or zinc white and copper carbonate, simultaneously with
nickel. So far, however, the results have not been attended
with success, inasmuch as the conditions requisite for nickel
precipitation, and more especially the necessity for maintaining
a high bath-temperature, cannot well be harmonised with those
requisite for the precipitation of good lead, zinc, or copper
pigments. He, however, succeeded in obtaining solutions of
* Zeitschri/t/ur Mektrochemie, 1900, vol. vii., pp. 21, 29.
Digitized by V^jOOQ IC
NICKEL.
295
lead chlorate in tlie anode cells without any of the lead passing
over into the cathode solution. The lead pigments desired
could then be obtained of high purity and good covering power
by treating the resulting solutions outside the electrolysis cells.
Again, the lead sulphate which deposits on the cell-walls in the
form of a felt-like coating may be readily converted into more
useful compounds of lead. The potential difference with a
current-density of 400 to 500 amperes per square metre of
cathode surface did not exceed 3*5 to 4 volts, although the
Fig. 170.
Fig 172.
![;u;le;lj;|1^^^^^ .^j,!!;!^ iji
Fig. 171.
Fig. 173.
Apparatus for the preparation of lead chlorate at the anode
and nickel at the cathode.
current-density was relatively high, and although clay cells were
used, and in spite of the deposit of lead sulphate on the latter.
Experiments are still proceeding in the direction of obtaining
other solutions in the anode cell.
Figs. 170, 171, 172, and 173 show a form of apparatus that
has been used for the above purpose. Clay cells are used as
diaphragms; they were obtained of suitable quality from the
Royal Porcelain Factory at Berlin, each about 525 mm. long,
525 mm. high, and 120 to 140 mm. wide [20J ins. x 20J ins. x
Digitized by V^jOO^ It^
296 ELECTKIC SMELTING AND REFINING.
4^ to 5^ ins.]. With the exception of these cells, however, the
apparatus, as illustrated in the sketches, exhibits no specially
novel points. Lead-covered wooden tanks are used, similar
to those employed in copper refining. These may be either
rectangular or triangular in vertical cross-section. The arrange*
ments for circulating the solution and for suspending the elec-
trodes are the same as for copper electrolysis.
Beview of Prooesses employed in Nickel Extraction. —
A review of the possibilities of treating nickel ores with special
reference to the experience gained within the last ten years will
suffice to show that electro-chemistry has been of great service
to the nickel industry, and that it is likely to be of further
assistance in the near future.
I. Boasting and Smelting Process in which no Provision
is made for the Separation from Nickel of any Copper
that may be present.
This process possesses the advantage that it is available for
the treatment of all nickel -bearing ores and metallurgical and
waste products. It consists of the following stages : —
Boasting of the Ores. — Sulphide and arsenical ores are
submitted to an incomplete oxidising roast in heaps, calciners,
shaft furnaces (pyrites burners), or reverberatory furnaces. The
products of the roasting are : —
(1) Sulphurous acid, which is available under some circum-
stances for the manufacture of sulphuric acid.
(2) Arsenic, which may be utilised for the preparation of
various arsenical compounds.
(3) ITie roasted material, consisting of oxides, salts, and
undecomposable oxides, sulphides, and arsenides of the metals
contained in the ore. It is further treated as follows : —
Smelting for Crude Matte. — The roasted ore (3) is sub-
mitted to a reducing process by fusion in blast furnaces (rarely
in reverberatories) with slags (8 and 15), refuse nickel products,
with oxide nickel ores, or with oxide nickel ore (especially
silicates, such as garnierite) mixed with sulphidising additions
(pyrites, black ash, or sulphates). This fusion results in : —
(4) Slags, which are rejected, or, in rare instances, may be
worked up into building stones.
(5) Metals, When bismuth and lead are present in the ores
these metals separate out in the metallic state; they contain
the precious metals present, and these latter may be separated
in the cupellation furnace.
(6) Crude nickel matte or nickel speise, which, in addition to
nickel sulphide, always contains iron sulphide or arsenides of
these metals, and, if the ores contain copper, copper sulphide.
Digitized by V^OO^ It!
NICKEL. 297
They usually contain 18 to 40 per cent, of nickel and copper,
taken together, and at least 40 per cent, of iron.
Roasting for Matte or Speise. — The raw matte or speise
(6), as also similar products from other works, are incompletely
roasted in the same way that the ore is treated in roasters,
reverberatories, or kilns; in this way gaseous and volatile
products are obtained, similar to (1) and (2), together with
(7) The roasted matte, which is qualitatively the same as the
residue numbered (3), but quantitatively is richer in nickel (and
in copper if that metal be present) and poorer in iron. This is
then treated by a
Matte or Speise Concentration Process, in which it is
finally submitted to an oxidising treatment in converters, after
reduction and fluxing, either apart from the last-named roasting
process using shaft furnaces, or in conjunction with that process
employing reverberatories or converters, substances rich in silica
being used in all cases. Of the gaseous or volatile products of
the roasting process ( 1) and (2), only arsenic (and that in part)
can be recovered from the flues and dust chambers. The other
products of the smelting process, lead and bismuth not being
present, are
(8) A dag rich in nickel^ which is used up again in melting
for crude matte, and
(9) Concentrated nuUte or speise; both of these contain but
little iron, but the matte, in addition to nickel sulphide, contains
as copper sulphide almost the whole of the copper present in
the charge. In treating copper-bearing arsenical ores of nickel
the bulk of the copper passes into a matte which may be
usefully worked up in the other matte processes. Matte and,
speise concentrated in this way are treated by
A Dead-roasting Process, in which they are submitted to
an oxidising roast, as complete as possible, usually in rever-
beratory furnaces. In this way they are freed almost entirely
from sulphur and arsenic, in the form of the gaseous and volatile
products named in (1) and (2), of which the latter (arsenic)
can be in part recovered from the flue dust. They are thus
converted into
(10) Nickel oxide, or, if copper be present, into
(11) Mixed oxides of nickel and copper. The last process to be
applied is then the
Reduction Process, by which the oxides are reduced.
Nickel oxide (10) free from copper is treated by an old process
which is still in use, in which it is mixed with rye or other
meal as a binding medium and with a refining medium (man-
ganese oxide, according to the patents of Basse and Selve),
and moulded into cubes, which are then packed in powdered
wood-charcoal and submitted to reduction by heating in mufile
furnaces. In this way the metal is obtained in the form of
Digitized by V^OOQ IC
398 ELECTRIC SMELTING AMD REFINING.
(12) Cyhe-nickdj which may be melted in crucibles, placed in
wind-furnaces or regenerative gas-furnaces, and run into the
form of either
(13) Granulated nickel, or
(14) Nickel bars or plates.
According to another process, which is specially applicable to
copper-bearing nickel oxides, the mixture of oxides is reduced
and fused with the addition of acid fluxes, either in crucibles or
in blast furnaces. In this way there are produced
(15) Slags containing nickel and copper, which may be worked
up by adding them to the charge in smelting for crude matte, and
(16) An alloy of copper and nickel. This alloy may have a
limited sale, as it is produced. The remainder is submitted to
Electrolysis in copper sulphate solution, of which the first
product is
!17) Electrolytic copper. Accompanying this is an impure
18) Nickel sulpTiate solution, which is purified by chemical
processes, effecting the precipitation of the last traces of copper
and the separation of iron, and is then electrolysed (preferably
recovering electro-chemical by-products at the anode), yielding
(19) Electrolytic nickel.
Any precious metals that may have escaped separation with
lead and bismuth (5) will be found in the
(20) Anode slimes which accompany (17), and which are
worked up in the lead and silver refinery.
II. Processes in which Copper and Nickel are Separated
during the Roasting and Smelting Stages.
This process has been found specially suitable for the treat-
ment of the Canadian magnetic pyrites (pyrrhotite) from the
Sudbury district of Ontario, containing nickel and copper.
This process, however, up to the stage at which the nickel is
refined by electrolysis (see below) has been in use for thirty or
forty years in the Hafod Isha Works of Messrs. H. H. Vivian
& Co. in Swansea.* It has nevertheless been patented recently.!
The first stages of the process are similar to those described
for Process I.
The Boasting of the Ore is conducted in heaps, the
products as in I. being —
(1) Sulphurous add, which, in view of the nature of the
roasting process used, can obviously not be utilised.
(2) The roasted ore (oxides, salts, iindecomi)osed sulphides,
and gangue stuff). This is now subjected to a process of
Smelting for Crude Matte in blast furnaces, yielding, as
in Process I.,
* Vaughan, Engineering and Mining Journal, 1883, vol. iv.
tGerman Patent 91,288, January II, 1893 (granted 1897).
Digitized by V^jOOQ IC
NICKEL. 299
(3) Slags, which are thrown away, and
(4) Crude matte, which, departing from the method employed
in the first process, is now
Smelted under reducing conditions in blast furnaces in
admixture with sodium sulphate and carbon, yielding
(5) Mixed matte, which separates into two layers in the slag-pots
before solidifying. These layers are of different specific gravity,
and can be readily separated mechanically by striking the block
after cooling. It thus breaks up into
(6) Tops, which consist of a copper-iron-sodium matte of low
specific gravity, and
(7) Bottoms, which are a matte of higher specific gravity and
rich in nickel. Since, however, the tops are not sufficiently free
from nickel, nor the bottoms from copper, both products are
re-smelted.
The Smelting of the Tops is conducted by smelting the
tops, (6) and (12), after they have been somewhat weathered [a
treatment hastened by sprinkling them with solution (21)],
mixed with crude matte (4) in blast furnaces under reducing
conditions. In this way there are produced
(8) Mixed matte, which is separated mechanically as before
[see (5)1 into
(9) Concentrated tops, which are poor in nickel, but rich in
copper, and
(10) Bottoms, which are mixed with bottoms (7) and submitted
to the process of
Smelting the Bottoms in blast furnaces with the addition of
sodium sulphate and carbon. This produces
(1 1^ Mixed matte, which is separated, after solidifying, into
(12) Tops, which are poor in nickel, but in which the copper
percentage must be increased by re-smelting with previous tops
(6). Meanwhile, however, the
n3) Concentrated bottoms are sufficiently poor in copper and
rich in nickel to warrant their being treated for the extraction
of the latter metal.
The Iieaching of the Concentrated Tops with water yields
(14) A solution of sodium compounds, which latter, after
evaporating off the water, are used in the treatment of crude
matte (4), and
(15) A residue, which consists chiefly of copper sulphide, but
contains almost all the silver and gold present. This residue is
then submitted to the well-known process of
Smelting for Copper, by which there is obtained
(16) Copper containing precious metals, which are separated by
Electrolysis, yielding finally
(17) Electrolytic copper,
(18) Silver, and
(19) Gold.
Digitized by VjOOQ IC
300 ELECTRIC SMELTING AND REFINING.
Chloridising Boasting of the Concentrated Bottoms in
reverberatory furnaces, followed by
Leaohing, have the effect of carrying into solution the copper
and part of the platinum group of metals. By precipitating
this solution with copper, there are obtained
(20) The platinum group of metals^ which are separated from
one another by recognised methods.
(21) Cuprie chloride solution, which is, in part, used to
hasten the weathering of the tops, (6) and (12), and is in part
precipitated by means of iron, from which is obtained
(22) Cement-copper.
(23) The residue from the leachiiig batlis consists mainly of
nickel oxide, with a little silica, sulphur, copper, iron, and
platinum. From this is obtained by
Beductlon in crucibles or blast fornaces, a
(24) Crude nickel mixed with the impurities named in (23),
which are separated by
Electrolysis in cyanide solution into
(25) Anode -mud, which is treated for the recovery of the
platinised metals.
(26) Fragments of ilie anodes, which are re-melted with the
crude nickel (24), and
(27) Electrolytic nickel.
CHAPTER III.
SILVER.
Occurrence in Nature. — Silver occurs native; or alloyed
with gold, copper, or mercury ; as chloride (horn silver), AgCl ;
as bromide and iodide in bromargyrite (AgBr) and in iodargyrite
(Agl) ; as a sulphide in silver glance, AgoS, both in a free state
and combined with other sulphides — e.g., red silver ores and
fahlerz — which may be regarded as silver thioantimonite and
thioarsenite respectively. As a sulphide, too, it occurs in more
or less noteworthy quantities in nearly all sulphide ores.
Properties of the Metal. — Silver (Ag; atomic weight = 108;
specific gravity = 10*5) is a white, highly lustrous, tough, and
malleable metal of crystalline structure (in the regular system)
and with a low degree of hardness, intermediate in this respect
between copper and gold. Its conductivity of heat and elec-
tricity is higher than that of any other metal. Its melting point
approximates to 1,000' C, and it is volatile at high temperatures,
and hence it may be distilled in the electric arc and before the
Digitized by V^OO^ It!
SILVER. 301
oxyhydrogen blowpipe. A special peculiarity ojE silver, and one
worthy of notice in connection with the refining and working of
the metal, is its capacity for dissolving oxygen when in a state
of fusion. At the moment of solidification the oxygen escaping
from the still liquid interior metal forces its way through the
superficial crust of solid metal, and often gives rise to a consider-
able loss of silver by the projection of small particles. The
phenomenon is known as the spitting, sproating, or vegetation
of silver.
Among metals which are soluble in, or dissolved by, silver^
special mention may be made of lead, mercury, copper, and zinc.
Silver is one of those metals which cannot be oxidised at high
or low temperatures, either in moist or dry air.
Of the metalloids, the halogens, and especially chlorine, have
the most active tendency to combine direct with silver ; and it
may also be united with sulphur by direct fusion. Among
sulphur compounds, hydrogen sulphide attacks this metal with
energy. The best chemical solvents are nitric acid and concen-
trate sulphuric acid. Various metallic chlorides (CuClg, HgCl.,,
and Fe^Clg) readily convert silver into its chloride, and thereby
render it soluble in other salts; and double salts, which are
soluble in water, are formed by the action of cyanides on metallic
silver, its haloid salts, or its sulphide.
Extraction. — In addition to the ores already mentioned, the
following waste or intermediate products from other manufac-
turing processes are utilised for the extraction of silver: —
Pyrites residues, regulus, slags, sweepings, and numerous alloys.
Chief among the latter are those derived from copper and lead
ores, and containii^ the precious metals, as black copper,
argentiferous lead, and the alloys of lead, zinc, and precious
metals obtained from argentiferous lead by concentration
processes.
The extraction is effected according to one of the following
principles : —
A. EXTRACTION OF THE CRUDE SILVER.
1. SOIiUTIOir OP THE SHiVEB IN ANOTHBK
METAIi.
Treatment with Lead. — This process consists in converting
the silver, in the ores or other raw material, into an alloy with
lead. When the silver content is low (less than 10 per cent.)
the ore, &c., is smelted along with lead ores ; but when the per-
centage of silver is higher, the material to be desilverised is
immersed in molten lead. In the former events it is necessary,
in order to minimise the loss of silver, to aim at producing a
Digitized by V^OO^ It!
302 ELECTRIC SMELTING AND REFINING.
silver-lead with not more than 1 per cent, of silver, and therefore
the lead treatment must be followed by one of the usual con-
centration processes (the Pattinson crystallisation process, or
alloying with zinc). The concentrated silver-lead alloy, or silver-
lead-zinc alloy thus obtained is afterwards separated into metallic
.silver and lead and zinc oxide (or metallic zinc) by oxidation
and distillation.
Attempts have been frequently made to effect the electrolytic
separation of these concentrated alloys, but, so far, without any
permanent success. As in these attempts it is not the silver,
but only the lead and zinc that take part in the electrolytic
reaction, the recital of the experience gained in this connection
will be relegated to the section dealing with the electro-chemistry
of those metals.
Amalgamation. — ^The precious metals are soluble in mercury
a.t the ordinary temperature, as they are in all readily fusible
metals. Silver need not even be in a free state in the ore, since
it can be extracted by mercury from all ores or metallurgical
products in which it is present as a haloid salt. Under suitable
working conditions, free silver sulphide also is acted on by
mercury. It is evident that in the treatment of ores containing
chlorides and sulphides, a suitable amount of mercury will be
required to convert these compounds, unless other metals, such
as copper, zinc, &c.y are specially added for this purpose. Hence
in the amalgamation treatment of ores and metallurgical products
containing such silver compounds, we have to do with an electro-
chemical process. The principles of this process, however, have
not yet been fully elucidated, although Krohnke, in describing
his well-known process for desilverising ores — which has been
extensively used in Chili, Bolivia, Peru, and Mexico since 1863
— expressly refers to the electro-chemical nature of the reactions
involved.*
Whilst no external source of electricity is drawn upon in the
Amalgamation methods based on the same principles as the
Krohnke process, numerous proposals have been made to utilise
■electrolysis to facilitate amalgamation in processes where this
reaction is concurrent with the preparatory treatment of the
ores. No practical results have, however, been obtained. In
order to avoid repetition, the most noteworthy processes of this
type will be described in dealing with the amalgamation of gold
ores, to which section the reader is referred.
The amalgam obtained by one or other of these methods is
separated by distillation into mercury and an impure silver.
Solntion of the Silver in Copper. — ^When ores containing
the precious metals and copper are treated by smelting processes,
a part of the silver and most of the gold are liable to pass into
the by-product that contains the copper, and, therefore, under
• B. Krohnke, Metholen zur EntsUhening von Erzefi, Stuttgart, 1900.
Digitized by VorOO^ It!
SILVER. 303
certain circumstances, into the copper itself. Ores containing
the precious metals are not thus smelted with copper ores, as
they might be with those of lead, in order to obtain the gold and
silver alloyed with the copper. Copper products containing the
precious metals are treated according to systems (2) and (3).
2. SEPARATION OF THE SILVEB BY PBOCESSES
OF CHEMICAIi SOLUTION.
The Ziervogel Process. — ^The matte smelted from the
argentiferous copper ores is so roasted that in the first stage,
among other products, copper sulphate is formed, whilst in the
second stage (after previous crushing of the roasted matte) a
double decomposition takes place between copper sulphate and
silver sulphide with the formation of silver sulphate. The latter
salt is then extracted from the roasted charge by means of hot
water and acid solutions of copper sulphate, and the silver is
finally precipitated from the resulting liquors by metallic
copper and iron.
The Aiigustin, Patera, and Kiss Processes. — The ores are
submitted to a chloridising roa.st, and the silver chloride is
extracted in the Augustin process by concentrated brine, in the
Patera method by sodium thiosulphate solution,^ and in the
Kiss modification by calcium thiosulphate solution. In the first
case the silver is precipitated in the metallic state by the use of
metallic copper, in the other processes it is obtained as sulphide
by the addition of soluble sulphides (Na^S or Ca(HS)2).
DietsePs Process. — ^A description of Dr. DietzeFs process
was given in the second German edition of this work based on
the patent specifications and private communications from him.
Dr. Dietzel has published additional particulars of his process, t
and as these data are undoubtedly based on recent experience
they will be employed in the following pages.
The approximate composition of jewellers' and silversmiths'
waste and the yellowish or bronze-coloured alloys recovered from
metallurgical products is as follows : —
Au, . .5-7 per cent. Zn, 8n, Pb, . about 5 per cent.
Ag, . . 22-60 „ Cd, Fe, Ni, \ ,^_^
Consequently, the main object is to recover the gold, silver, and
copper in the metallic state ; and this is effected by the aid of an
electrolyte consisting of a weakly add solution of copper nitrate.
* [This salt is still commonly known as hyposulphite of soda, but as this
term belongs more properly to a different compound alto^ther, the name
which is required by the modem system of nomenclature is adopted in the
text. —Translator. ]
tZ./. Elektroehemie, 1899-1900, vi., 81.
Digitized by VjOOQ IC
304
ELECTRIC SMELTING AND REFINING.
The copper and silver are dissolved at the anode; the gold
remains undissolved ; the copper being then deposited at the
cathode and the silver outside the cell.
Solution run off.
Fig. 174.
The arrangement of the electrodes and supplementary appa-
ratus is shown in Fig. 174, which represents a cross-section of
the dissolving vessel. KK are the rotary, cylindrical, copper
cathodes, coated with a thin layer of grease or graphite, on
which the deposition of copper takes place. As soon as the
deposit assumes the dendritic form, it is knocked oflT, and, as the
copper contains a little oxygen, it is smelted with charcoal. The
copper cylinders are suspended on flanged contact rollers which
are set in motion, the rolling friction between the two surfaces
causing the cylinders to rotate. This method presents the ad-
vantage that the driving mechanism and shafts are out of contact
with the liquor, while the bearings can be oiled and used as
conductors for the cathode current. P is a loose bottom for
supporting the material to be treated, S, and is constructed of
hard rubber, celluloid, or glass plates set in wooden frames. K
necessary, the bottom can be arranged to run on porcelain rollers,
r, and rails, 8, The plates, P, are provided with platinum wires,
or with some insoluble conductive material sudi as plates of
resistant carbon, or even strips of chromium, forming the ter-
minals of the insulated line wires. That chromium can play the
part of a precious metal has already been shown by Hittorf.
These inverted roof-shaped bottom plBites are well covered with
the material to be treated, which is now usually cast in the form
Digitized by VjOO^ li:!
SILVBB. 305
of easily-handled plates, ^ to | inch thick. Owing to the rela^
tive position of the electrodes being more horizontal than vertical^
the utilisation of space in this cell is less advantageous than in
other t3rpes. An anode base will not hold much more than 5 to
8 kilos, of auriferous silver material, or at most I cwt. of material
to be treated, on a space of about 20 square feet, and the opera-
tion takes about six days to complete. D D are linen ftlter-cloths
forming the bottom of the cathode chamber. In the new plant
there are a number of these loosely suspended in series in a long
dissolving vessel, in such a manner that enough space is left for
removing the anode carriers, and that a cathode chamber is in
position over each anode frame. The object of these filters is to
catch any dendritic copper falling from the cylinders, and to
prevent the anodic silver solution from ascending to the copper
cathodes as a result of movement set up in the liquor by the act
of removing the gold sediment or through the disengagement of
oxygen. Above the cylinders, K, the desilverised electrolyte
drops into the cathode chamber, traverses the linen filter, and,
after dissolving copper, silver, and a few accessory constituents
from the charge of material, runs away through slots on a level
with the top of the frame into a glass pipe provided with lateral
perforations. This pipe is connected with an adjustable regula-
tor, so that the liquor is maintained at a constant level. Any
gold sediment accidentally carried away is retained by simple
filters. The small quantity of sediment washed from the anode
frames is cleared out of the bottom of the vessel once a year.
The dissolved silver is brought into contact with old copper in
vessels set up on a lower level, so that eventually the whole of
the silver is recovered in the crystalline form, and an equivalent
amount of copper is returned to the liquid.
As can be seen from the table, the ratio of dissolved silver to
dissolved copper is approximately = 1 : r5 by weight. Since
the silver equivalent is 3*4 times greater than that of copper, it
follows that, when the silver is displaced by copper in an inter-
mediate operation, the consumption oi the latter will be in the
ratio ^-. : 1*5, or, in round numbers, 20 per cent, of the weight
of copper in the alloy, or 7 per cent, of the weight of the entire
alloy. According to the experiments of Le Blanc and others, the
decomposition of the copper salt necessitates a tension of 0*44
volt higher than is needed for the silver salt. In consequence of
this great difference, the tension of the bath has to be about 0*5
volt higher than if the process were arranged for the cathodic
deposition of silver from the same anode material. This draw-
back is, however, insignificant in comparison with the advantage
mentioned, namely, that the silver can be rapidly and quantita-
tively deposited by means of copper, outside the cell and without
the aid of any supplementary current. With a little attention
20
Digitized by VotOOQIC
306 ELECTRIC SMELTING AND REFINING.
to the circulation of the electrolyte, the silver is all deposited
before the last precipitation vessel but one is reached. Tliis can
be demonstrated by the hydrochloric acid test. Nevertheless, it
has not so far been possible to prevent the small loss of silver
finally traceable in the granulated copper. In practical working,
one must allow for a loss of 0*03 per cent, of silver, though, in
the author's own experiments, the average silver content of silver
in the copper was only 1 gramme in 10 kilos.
In its passage from one vessel to another, the liquor finally
becomes almost neutral, and when this condition is reached the
dissolved iron is deposited as a basic ferric salt Should the
material under treatment contain much iron, the latter is
precipitated, almost quantitatively, in an aerating vessel (not
shown in the Fig.) provided with beaters, and is retained by a
filter. The liquor next flows into a collecting tank, and is
pumped thence, or forced by compressed air, into a storage tank.
Here it is slightly acidified with nitric acid and returned through
glass pipes into the cathode chambers. The percentage of copper
in the solution ranges between 2 and 5 per cent., and that of tree
nitric acid between 0*05 and 4 per cent. The average current is
about 150 amperes per square metre, with a mean tension of 2^
to 3 volts in the bath. A specifically heavy mixture of gold and
silver, generally accompanied by traces of copper, remains behind
on the anode carriera. The residues are dark brown to deep
black in colour, and of granular to amorphous structure. The
successful exhaustion of the raw material in this process is due
to the fact that it is supported by a conductive surface instead of
being freely suspended. Odd pieces of thick imperfectly-corroded
alloy are included in the succeeding charge. The sediment also
contains lead (as peroxide) and tin (as basic nitrate or stannic
acid), whilst zinc, cadmium, nickel, iron, &c., pass into solution,
where they mostly remain (except the iron). The gold is
recovered in a pure state by boiling the sediment in dilute nitric
acid, dissolving the residue in nitro-hydrochloric acid, filtering
off Uie lead chloride, silver chloride, and stannic acid, and finally
precipitating the metal with ferrous chloride. Platinum is
thrown down by iron in the process of recuperating the iron
solution.
According to earlier reports, the difficulty of separation is
increased in proportion to the amount of associated metals that
are not precipitable at the cathode, but become insoluble at the
anode. Of these, zinc in particular consumes an appreciable
fraction of the power in the anode currents, and causes an
equivalent precipitation of copper. This impoverishment of the
electrolyte may be counteracted by slightly increasing the acidity
of the store of desilverised liquor beyond the degree otherwise
most advantageous — ^namely, to an average of not less than 0*2
per cent. Under these circumstances, copper is dissolved at the
Digitized by V^jOOQ IC
SILVER. 307
cathode, and in the precipitating vessels after the silver has been
deposited. Moreover, the percentage of nitrous acid in the
liquor is diminished, the result being to retard the cathodic
precipitation of copper to such an extent that it does not exceed
the amount introduced at the anode, and the electrolyte is not
impoverished in this metal. It is worthy of note here that the
presence of copper nitrate retards the reduction of nitric acid to
ammonia; at least, no ammonia can be detected in the liquor
with the low amperage here in question. At the end of a year
the liquor is so rich in zinc and other non-precipitable metals
that it must be renewed, the old liquor being then worked up.
So long as the associated metals, especially lead and tin, do
not exce^ a small percentage, the separation of the three
principal components — gold, silver, and copper — ^proceeds with
such uniformity that the gain in copper not only covers the cost
of the process, but even yields a profit, whereas, a few years ago,
this metal was regarded as a troublesome factor. Even with
very difficult alloys, a daily quantity of only 130 lbs., and expen-
sive steam, the cost of separation does not exceed the usual
average of about 5d. per lb., without making any allowance for
the gain of copper. The richer the alloys in silver, and the
purer they are, the more profitable the separation, since, in con-
formity with its high equivalent, silver is deposited 3*4 times as
quickly as copper. The dry metiiod for gold always ensures a
high profit^ so that the cost of refining the gold sediment need
not be considered. And it is evident that the profits will
increase out of all proportion the larger the daily output and the
better the utilisation of the available power.
Andre * experimented on the separation of similar alloys long
before Dietzel, and also worked on the principle of dissolving
copper and silver at the anode, precipitating the silver on copper,
and finally recovering the copper at the cathodes of the electro-
lytic apparatus. Most of the reports on Andre's patent in the
technical press merely repeat the wording of the specification as
to the separation of copper and nickel. With regard to the
separation of gold, silver, and copper, he makes the following
interesting statements : —
*' The alloys or semi-manufactured products in question, in the
form of granules or plates with clamps, are connected to the
conductor and suspended, as anodes, in a bath of dilute (1:10)
.sulphuric acid. Between the anode and cathode is placed a
frame, which fits perfectly tight against the wails of the vessel,
and is covered on either side with cotton fabric or a porous
membrane. In this way the bath is divided into two chambers,
one for the anodes, the other for the cathodes ; whilst the frame
itself is filled, between the membranes, with copper scraps,
granules, Ac. Under the influence of the curreni^ silver and
• German Patent 6,048, Nov. 1, 1877.
Digitized by LjOOQ IC
308 ELECTRIC SMELTING AND REFINING.
copper are dissolved at the anode, whilst the gold and other
admixtures remain behind. In passing through the frame filled
with copper, on the way to the cathode, the silver is deposited,
and the copper solution arriving at the cathode deposits pure
metallic copper there. The cathode may consist of carbon^
copper, or other suitable substance. When the granulated
copper in the frame is almost completely coated with silver, the
frame is recharged with copper, the cement-silver adhering to
the granules being washed off and cupellated.
" This method of separation reduces the consumption of acid
to a minimum, saves labour, and furnishes chemically pure
copper (very suitable for electrical purposes) instead of an
inconvenient mass of copper sulphate.
"The arrangement may be modified as follows, without de-
parting from the principle of the process : —
"The cathodes are laid flat or aslant on the bottom of the
vessel, and over them in another vessel come the anodes (in
granular or plate form). The bottom of this second vessel is
formed by a membrane of low porosity, or preferably by two
such membranes with a small free space between.
" Here again the current causes a solution of silver and copper
to form at the anode, which solution is uninterruptedly removed
at the bottom of the vessel and replaced by fresh, dilute acid
entering at the top. The silver and copper liquor is run into a
vessel below, where the silver is deposited on granulated copper.
From this vessel the desilverised liquor passes into another
vessel, and is freed from copper at the ca^ode. Finally, the
liquor is gradually drawn off from this vessel into a storage tank
below, and is returned to the anode for use over again in dis-
solving silver and copper there. The plant may be arranged
terrace-wise or mounted in a tower, according to local circum-
stances. To prevent diffusion, the liquor between the two
membranes is run off, slowly but continuously, and re-united with
the silver and copper liquor. Whether the terrace be high or
low, the lowest level always contains three collecting tanks : —
No. 1 to collect the effluent liquor. No. 2 for the desilverised
copper solution, and No. 3 for the decopperised liquor (dilute
sulphuric acid). Under each tank is a pressure vessel, from
which the liquors can be forced by compressed air up into the
three corresponding tanks on the uppermost terrace.
" The vessels may be of stoneware or wood (petroleum barrels
do very well), and the pressure vessels may be of wood, with
iron hoops, provided the pressure required does not exceed that
of the atmosphere. For higher terraces, small, lead -lined
copper cylinders are used."
Digitized by VjOOQ IC
SILVER. 309
3. SOLUTION dF THE CONSTITUENTS OTHEB
THAN SHiVEB.
WohlwiU's Process. — Besides the Freiberg process for the
treatment of copper mattes by sulphuric acid, the Harz sulphuric
acid extraction process for black copper containing the precious
metals, and the Bossier process, by which the metals other than
silver are removed as sulphides by a smelting process, electrolytic
methods have now found an extended use. It will suffice to
refer to the electrolytic copper refining, described at length on
pp. 187 to 272 ; as also to the electrolytic treatment of crude
nickel and of nickel-copper alloys, explained on pp. 1^75 to 296.
Experiments on the electrolytic treatment of zinc skimmings
and of argentiferous lead will also be described in the chapters
dealing with zinc and lead respectively.
The process introduced by Wohlwill for the treatment of
copper alloys rich in silver after the manner of the electrolytic
refinery, and with the use of copper sulphate as electrolyte, is
worthy of special consideration, contrasting, as it does, with
the Dietzel process, in which the silver and the copper are
carried into solution together. Wohlwill succeeded in so
applying his process that the silver remained quite undissolved
even when it was present to the extent of alx>ut 30 per cent.
The following remarks were privately communicated by him
to Dr. Borchers, with a verbal permission for their publication
in this work : —
''Since the whole of a small quantity of gold, silver, and
platinum present in black copper may be recovered by the
electrolytic process, the question arises, to what maximum limit
may the proportion of precious metals in copper alloys extend
witiiout rendering their separation by such a method impossible.
I [Wohlwill] have had an opportunity of dealing with this
problem on a practical scale, since, in the years 1877-78, the
small silver coinage of the German Confederated States was
called in, and was, for the most part, sent to the NorddeiUsche
Ajfflnerie for treatment. By the method commonly employed for
the recovery of the silver from alloys of this kind, the copper is
obtained as sulphate. But, in view of the necessity to treat
large quantities of the material in a short space of time, the
dinsct recovery of metallic copper, such as is possible by electro-
lytic treatment, appeared to offer special advantages. Experiment
showed that, in practice, the silver alloy, rich in copper, when
used as an electrode material for the electrolysis of copper
sulphate, differed from argentiferous black copper anodes, only
in that a sponge rich in silver remained adhering to the anode
after the copper had been extracted from the surface of the
metal, and that the sponge covered up that part of the alloy
Digitized by VorOO^ It!
310 ELECTRIC SMELTING AND REFINING.
which was below the surface, and so checked further action;
moreover, on continuing the electrolysis, some of the silver
passed into the solution, and in a short time the anode became
covered with difficultly-soluble silver sulphate. The natural way
of obviating this difficulty was to remove the silver coating by
mechanical means, to submit the cleaned plate once more to
electrolysis, and to continue the rubbing and electrolytic treat-
ments alternately until it was found that the anode could no
longer bear the mechanical treatment. The greater part of the
small German coins sent to the Hamburg Institute for separation
— namely, the silver groschen, 3- and 6-kreuzer pieces, schillings,
d:c. — was, at that time, submitted to this process, after having
been cast into plates.
" It is evident that the cost of the ordinary electrolytic process
must be increased considerably by the hand-labour necessitated
to a considerable extent in a process of this kind. It was, there-
fore, necessary to study the conditions, and to endeavour to
secure as full and as long-continued an electrolytic action as
possible, without at the same time permitting any of the silver
to pass into solution.
" If even a very small quantity of silver should pass into the
solution with the copper, difficulties would arise, not only from
the liquid immediately becoming milky, but also from the
.properties of the deposited copper becoming altered. In that
case, instead of a dense adhesive and reguline copper deposit,
there would be thrown down a flocculent clear red precipitate,
containing silver as well as copper ; and this would form a spongy
coating on the cathode. The operation of a bath that has so
deteriorated must be suspended until every trace of silver has
been removed from the solution. In order to obviate this
difficulty, as also to prolong as much as possible the period during
which the copper is dissolved and precipitated, it was found
necessary to apply a lower current-density and to use a weaker
solution of copper sulphate than is employed in the refining of
black copper. But, above everything, it was essential to the
good working of the process that freshly-scraped anodes should
be used side by side with those on which there was a more or less
heavy coating of silver. Hence the anodes of every tank were
divided into two or three groups, each of which could, in its
turn, be submitted to the process of mechanical cleansing of the
surface. Under these circumstances the current always divided
itself between the anode plates of each bath in such a way that
those on which the silver deposit was thinnest became subjected
to the strongest attack, whilst even with those on whicli the
deposit was thickest the electrolytic solution of the copper
proceeded gradually to an increasing depth without any of the
silver being attacked. By adopting such an arrangement,
therefore, the extraction of the copper could, on the whole, be
Digitized by LjOOQ IC
SILVER. 311
pressed farther than by the simultaneous treatment of similar
anode-plates.
"Alloys containing upwards of 30 per cent, of silver could
thus be advantageously treated. As the proportion of silver is
increased the period at which the spongy silver must be mechani-
cally 
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Open Library
American Libraries
TV News
Understanding 9/11